This research documents how we use our FDV prediction tool on the Espresso ($ESP) token during the first 48 hours after TGE. The goal is not to claim perfect predictions, but to generate a tradable range (P10/P50/P90) and a clear diagnosis of what the market learned between TGE, day 1, and day 2.

FDV Market Values

In the first 48 hours, FDV behaved more like impact discovery than fundamental assessment, with the market valuing:

1. Activation of sellable inventory

   Claims, farmers, hedgers, and early holders when they become active are just as important as how much there is.

2. Effective depth (not headline depth)

   Depth is reduced by fragmentation and coverage of offers. The important number is: how much can absorb the flow under pressure?

3. Stabilizers (increased MM/inventory support)

   Stabilizers don’t appear instantly; they increase gradually. The time until support changes the entire distribution.

The result is Initial FDV is naturally distributional, which is why pre-market crypto FDV should consider ranges and probabilities, not a single deterministic target.

The Model We Use

We view the launch as a system in which price movements are influenced by volatility plus the impact of flows into market depth, adjusted by stabilizers:

Explanation:

• NetFlow(t) = how much net sales appear, when

• EffDepth(t) = how much real liquidity remains after discounting

• Stabilizers(t) = how quickly the market structure improves

• Output = FDV distribution over time: P10 / P50 / P90

These percentiles form the basis for translating launches into probability views.

Espresso (ESP)

After importing data on Supply and Emissions data, a comparison is made with direct competitors. The goal is to estimate the FDV listing using competitors as a benchmark. As a retail trader, this component is quite useful for building bias. Espresso’s direct competitors here are Celestia, Avail, EigenCloud, and Metis. Using the High FDV scenario and % projected FDV & Market Cap 100% as benchmarks taken from Espresso’s utility performance before TGE, the estimated FDV for $ESP is $221.16M and the market cap is $33.3M, missing ~20% of the actual TGE data.

Next, we make predictions using the tools we have built. The settings used are:

• Daily volatility 15%

• Max drawdown 55%

• Liquidity Depth Tier Normal with pair type Stable (ESP/USDC)

• Scenario Base and Market maker active

Our model estimates that the best FDV within a 48-hour range is $221.16M and the worst FDV is in the range of 99.52M - 110.26M, which is still relevant to the actual FDV after TGE.

What matters for an FDV market:

• Day-1 FDV ($208.45M) lands inside the 24h band, slightly below the median.

• Day-2 FDV ($219.72M) lands very close to the 48h median.

That is exactly what you want from a market-usable forecast: not “nailing the print,” but placing reality inside a credible band and identifying where settlement happens.

Insight

1. Don’t trade too often at the opening price

The TGE opening price may be a temporary result generated by:

• incomplete seller activation

• thin/fragmented market depth

• stabilizers not yet fully active

In an FDV market, this means the initial price may reflect “temporary conditions” rather than the 24–48 hour settlement regime.

2. The 24-hour checkpoint is when flows are revealed

Day 1 landed within range but below median consistent with:

• delayed net sales becoming visible

• liquidity discount (effective depth < expected depth)

This is the period where Polymarket pricing often shifts from narrative to mechanism.

3. The 48-hour checkpoint is where settlement becomes measurable

Day 2 FDV Espresso approaching the P50 model indicates the market has found a stable absorption zone.

For Polymarket, this is the most useful horizon: settlement tends to be more stable and modelable than the first print.

Closing

For the Polymarket FDV market, the advantage lies not in predicting the first opening price, but in predicting where the launch regime will stabilize once flow, depth, and stabilizers become observable. In Espresso ($ESP), the results for day 1 and day 2 clearly fall within the model’s limits, and the 48-hour settlement approaches the median regime. That is the right structure for turning an uncertain launch into a tradable probability.

• Market always overreacts at the beginning, especially markets with tokenomics docs that have not been published. Generally, there will be consistent downward revisions → the average maximum FDV expectation is between $300-500 million.

• Event information such as tokenomics documents, airdrops, and TGE act as the main validity odds and usually accelerate the decline in probability into actual FDV.

Core Pattern: Hype Decay

• Starts optimistic (high MC probs 0.5–0.9)

• Ends very conservative (high MC probs near 0, low MC probs near 1)

• Average expected MC drops from $500M–$3B → $100–500M

• Strong negative time correlation for high thresholds (–0.6 to –0.88)

• Low thresholds sometimes rise slightly toward certainty

Event Impact (average changes)

Events explain ~40–50% of the large probability moves.

Tokenomics Release – Two Clear Behaviors

Early release (Day 1 / before market opens)

• 5 markets (HeyElsa, Kinetiq, Seeker, Immunefi, Sentient)

• Starts more realistic (lower high-MC probs from day 1)

• Lower volatility throughout

• Still decays gradually → early info prevents wild hype but doesn’t stop the overall downtrend

Late release (mid-market)

• 4 markets (Lighter, Infinex, Fogo, Stable)

• Pre-release: inflated high-MC probabilities

• Release moment: small to moderate downward adjustment (especially >$500M–$1B)

• After release: decay usually accelerates slightly

• Net: corrects over-optimism built during “no info” phase

Project Groups – Hidden Differences

Short-lived markets (e.g. HeyElsa ~3 days) show extremely compressed hype cycles.

Takeaways

• High FDV bets usually lose value steadily → strongest decay on >$1B thresholds

• Tokenomics release = mild-to-moderate sell signal for moonshot positions

• Airdrop + TGE = stronger sell signals

• Safest area in most cases: <$200–500M thresholds (final probabilities often >0.9)

• Prediction markets appear reasonably efficient at incorporating new information

“Note: I first published this as an X post. This version is updated and edited into a proper research note.”

TL;DR

• Sports: prices tend to correct fast (≈20 hours), but outcomes remain high-variance.

• Non-sports: price discovery is slower (≈20–90 hours), consistent with slower information arrival + weaker consensus formation.

• Across both: “obvious mispricing” isn’t automatically an edge; extreme-price conviction shows weak results and looks more like repeated overconfidence than pure luck.

This analysis uses Polymarket trades data that filtered to focus on moments where price move and liquidity are meaningful, not a small and dust trades.

Filters:

• Only trades ≥ $1,200 as a liquidity proxy at the moment of entry

• Trades associated with high-performing trader stats

• Markets split into sports vs. non-sports

Efficiency ≠ predictability

Basically, we look at a prediction market price as a noisy estimate of probability, we always asking:

1. How long does the market take to incorporate information after a trade (correction time)?

2. How does price move after entry (path dependence / drift / mean reversion)?

3. Do extreme entries behave like edge or like overconfidence?

A good way to answer it (technically):

• Efficiency: price rapidly reflects available information (low predictable alpha after public info).

• Predictability: ability to forecast the underlying event outcome (not the same as efficiency).

That’s why “efficient market” can still produce a lot of losing trades. Don’t treat efficient into “easy money”. Treat it into the obvious stuff gets corrected fast.

Mispricing Exists, But it’s Often Misdirected

Big mispricing does show up. The problem is, it doesn’t reliably point toward a clean edge. A lot of the time, the “wow that price is wrong” feeling is just confidence talking.

In other words, mispricing magnitude alone isn’t a thesis. We still need information advantage, timing advantage, or structure advantage.

Mispricing Distributions: Sports vs Non-Sports

• Sports mispricing clusters more around the middle range (the market “snaps back” fast when information hits).

• Non-sports shows more persistent small mispricings (looks like slower, messier information flow).

“non-sports behaves like a market where consensus formation is staged, you don’t get one clean jump to “truth”, you get stabilization into re-stabilization as more signal arrives.”

Correction Speed: Sports Correct Fast, Non-Sports Drift Longer

• Sports: most markets correct within about ~20 hours

• Non-sports: often sit mispriced for ~20–90 hours

Microstructure interpretation:

• Sports has faster information arrival, more participants, and often external reference pricing (odds, news cadence, etc.). That accelerates price discovery.

• Non-sports often has slower signal, fuzzier fundamentals, fewer hard anchors, and more disagreement. That stretches correction windows.

Also, sports outliers exist because some traders take positions way before a match starts, when uncertainty is still huge.

“Whale Correctness” Maps: Signal vs. Confidence at The Tails

These plots below are basically asking; when a large trader shows up, does that trade align with eventual resolution, and how does that relate to mispricing?

You can see clusters where large trades align with resolution and clusters where they don’t. The presence of “whales” is not automatically a cheat code. Sometimes it’s informed flow. Sometimes it’s just big conviction.

“Simply just don treat big size as smart size.”

Wrong-Side Rate: Sports Can Be Efficient and Still Punch You in The Face

~33% of sports entries are on the wrong side. That doesn’t mean the traders are bad. It’s consistent with the idea that:

• sports pricing can be efficient, and

• outcomes can still be intrinsically high-variance.

“Efficiency doesn’t remove randomness from reality”.

Mispricing + Time Together Shows a Real Structure Difference

This is where the split becomes most actionable as understanding, not signals.

• Non-sports can remain wrong for a long time even when mispricing is extreme.

• Sports can deviate, but the market usually corrects faster once information hits.

These charts help visualize how “mispricing intensity” relates to your correction/behavior metric. Even without turning it into a trading rule, it shows the markets behave differently under stress.

Conclusion

Overall, the data points to a clean structure split, sports markets behave like faster price-discovery venues where public information gets absorbed quickly, but outcomes remain noisy enough that “efficient” still feels brutal;

non-sports markets behave like slower consensus machines where information arrives unevenly, so mispricing can linger and price can drift for much longer. And in both categories, the recurring lesson is the same extreme prices and strong conviction don’t reliably translate into edge, and a lot of the bad performance at the tails looks more like systematic overconfidence than random bad luck.

We are paying science the wrong way. The problem that all scientists face is an unfair and non-transparent reward and incentive system. A series of papers in 2014 in the Lancet, for example, estimated that 85 percent of investment in biomedical research is wasted. Many other disciplines have similar problems.

In 2022, the US spent $923 billion on R&D, far more than China's $812 billion. But in the context of science, China occupies the first position in terms of its share of cited papers. This shows that large funding does not have a big impact on the continuation of discoveries or improvements that can be made.

Decentralize Science (DeSci) is here to create a system where scientific innovation becomes permissionsless and unstoppable. Being able to open up opportunities for everyone to access science that is “objectively true” and the valuable ones should be funded. DeSci focuses on decentralizing how scientific information, data, and IP are shared, accessed, and funded. TradSci challenges:

• cost and IP; healthcare & drug costs are very expensive, partly because of the old patent system. so pharmaceutical companies can control prices.

• Funding challenges; gov funding such as NIH grants can be inefficient and political, and industry funding prioritizes care over treatment or only focuses on areas that are profitable, not necessarily the best for patients or society

• Permission system; innovation is still fully controlled by industry needs and highly dependent on the government.

House of All Leading Biotechnology DAO

Bio Protokol is designed to support potential biotechnology to be more efficient, open, and market-driven from stage funding to liquidity. The launchpad mechanism is similar to other launchpads, $BIO holders need to stake (Bond) to get access to liquidity engines, incentives, BioAgent and other eco services.

• $BIO token holders curate which communities and projects to launch using DAOs mechanism.

• then DAO raises funds from the public using the bonding curve method. the money raised from the bonding curve will go to DAO to finance operations and research.

• Initially bonded tokens will be used together with new DAO tokens to create liquidity (tradeable). Bio Protocol holds 50% of the initial liquidity pool

• Bio Protocol also holds a 6.9% share of each launched DAO total token supply. meaning Bio protocol holds all DAO tokens launched in the ecosystem

• BIO earns revenue from trading fees on the bonding curve and fees from liquidity. this allows bio protocol to cover operational costs from its own activities. rather than having to issue more tokens or raise funds from outside.

$BIO Token Value & Revenue

token is used for governance over the dao ip portfolio. Value comes from the expectation that the IP will become valuable over time.

• Monetization: Dows could generate revenue by selling products developed from their research (hair dao selling shampoo). Revenue could be used for things like buying back and burning the dao tokens, which might reduce supply and potentially increase value for holders.

• IP value: successful research can lead to valuable IP that can be acquired by large companies (Pfizer investing in and potentially being interested in IP from Vibe Bio funded projects) if the IP is sold. The decision to distribute the value to holders is decided by the community, but direct payouts can raise complex legal issues such as KYC or securities regulations.

• market forces: like traditional biotech stocks, token prices can reflect the perceived success and potential of the DAO scientific projects. good data and progress can increase value.

When the DAO Token was launched, a 6.9% share of each DAO's FDV was turned into a source of revenue. DAOs that achieve bonding also provide $BIO owns 50% of initial liquidity pools. In addition, trading fees from the bonding curve will be a source of revenue for BIO Protocol. This revenue is used by the protocol to cover operational costs, rather than issuing new tokens or seeking funding. (dashboard as of May 30, 2025)

Catalyst Growth significant for Bio Protocol

There are several catalysts that can bring good growth to the Bio Protocol in the future. Some of these relate to the utility and mechanics of the protocol:

BIOPSY-22

Bio Protocol team initiated BIOPSY-22 proposal, the team wants to change the vesting system that will unlock on May 28, 2025. They want to change the 6 years token vesting (1 year cliff) to linear 6 months (14 nov 2025 - 14 may 2026) instead of full unlock on the 28th. The remaining 5/6 will be linear vesting for the next 5 years (token emission per year ≈117.3M BIO per year).

This decision was taken as a strategic step to keep circulating supply low during the mechanism development protocol revamp process. Of course, this decision will keep the dump pressure token stable.

BIO V1 & Chain Expansion

The first major protocol upgrade was done to re-design how biotech is more efficient, transparent, and scalable. The main features are Permissionsless Launchpad, Automated Liquidity & incentive engine, and BioAgents. Coupled with chain expansion starting from Solana, the goal is to make the protocol more accessible to all DAO customized use cases.

$BIO holders will determine which projects are worth launching and get rewards from early allocation. Once a project is approved, it automatically gets liquidity + incentives for scientific milestones. AI agents are also automatically created for core scientific tasks to cut traditional costs.

BIO AI Scientist

Evolving BIO into a new layer for scientific capital supported by agentic ai. BIO liquidity is able to provide 3 main parts:

• commuities (DAOs) - status quo

• Researchers (IPTs) - coming soon

• Agent builders (AI scientists) - coming soon

Agentic BioDAOs are AI swarms that generate onchain IP to help global research commuities. Instead of monthly negotiations with universities, science, experimental plans and IP can be natively generated on chain. So each DAO that launches in the Bio Ecosystem will get a custom agent using the community's data in the hope that the community's data contribution will grow more efficiently.

Upcoming Launch

• NootropicsDAO — create open-science ecosystem of solution for human cognitive enhancement that aligns professional researchers, citizen-scientists, and biohackers.

• MicrobiomeDAO — decentralized network funding and translating microbiome science into real-world health solutions.

• KidneyDAO — funds early diagnostics, novel therapies, and collaborative data systems to reduce dialysis dependence and improve kidney health outcomes worldwide.

• SleepDAO — bridge the gap between cutting-edge sleep research and practical solutions, empowering individuals to experience 20% more energy, productivity, and performance.

• GlioDAO — global movement using blockchain to fund cutting-edge brain cancer research.

• StemDAO — change stem cell therapy yang sudah hostage for too long, buried under bureaucracy, patents, and pharma lobbying.

$BIO Risk & Advantages

• Narrative Mindshare — DeSci is not a leading narrative at the moment, but it is possible that $BIO revamp in the use of AI agents could have an impact on the project's mindshare and narrative. (data as of May 30, 2025)

• Price Action & Volume — $BIO perps volume on CEX is ranging at $22M - $34M. If the volume can stay in that range and the price does not drop to $0.05 then it is still a good time to accumulate with 60% potential upside at $0.135.

Final Thoughts

We are bullish about the development of BIO Protocol which is now starting to follow the flow of liquidity. With the concept of fair launch design, $BIO has become the bet of all bioscience projects. DeSci projects listed on crypto data aggregators such as coinmarketcap or coingecko are dominated by projects that launch using Bio Protocol.

We also think that Bio Protocol will still be the market leader bioscience launchpad in the future, with the integration of BioAgent making $BIO even more attractive for liquidity to enter the ecosystem. The plan to hold and change the unlock mechanism is also in line with the future potential of the protocol upgrade plan. Maintain the fundamental value and fair mechanism in paying for science.

• Market Potential: Positioned in the growing generative AI chipset market with projected 32% CAGR by 2030, competing with established players like RENDER and Akash Network.

• Unique Value Proposition: Combines DePIN, DeAI, and AI Agent narratives with a focus on providing decentralized GPU resources and supporting complex AI agent development.

• Heurist's Infrastructure Scale: Currently operates with +14,000 active GPUs, 511,608 GB total VRAM, and over 22,000 cumulative miners, demonstrating robust network capacity.

• Dynamic Emission Model: Features a 3-phase pre-mainnet mining reward system and adapts emission rates (1.25%-5% annually) based on network demand.

• Agent Framework Innovation: Provides a lightweight, Python-based framework for building AI agents, focusing on ease of use and integration with decentralized protocols.

Mindshare about AI is on the rise these days. the power of AI is considered capable of eliminating some repetitive human work. Is it happening now? YES, it is already happening. It's an inevitable fact that AI has started to coexist with humans. This condition will not be seasonal, it will happen until technology consumes human life completely.

Where are we now? At the end of wave 2 imho. But far before that, the world has been helped by “Technical Computer” things. Starting from rule-based that simplifies the process of repetitive work and/or those that require logic functions. The wave got bigger when the development of semiconductors began to be useful for improving computer performance, the result was that theories about mathematical statistics, physics, and ideas about Artificial Intelligence (AI) could be simulated. This wave continues to move until AI can be used only based on trained data (historical data), until AI is able to generate according to the input given. This development marks the end of the wave and leads to a new wave with higher acceleration of Generative AI and computing.

Each stage of technological development, there are definitely issues that need to be improved continuously. One of them is AI infrastructure, either software or hardware to run Generative AI. Some of them are GPU, Virtual RAM, Storage, Data, to electrical power. Many large companies are developing supporting sectors of AI infrastructure, one of which is in the semiconductor industry. An important component for all electronic devices, NVIDIA as one of the semiconductor companies continues to get attention from product developments that continue to increase rapidly.

In 2023, NVIDIA is in the top 5 semiconductor companies. Following the success of OpenAI and the need for chips to run Advanced AI. This need will continue to exist following the development of increasingly complex AI infrastructure. Entering wave 3 from cognitive AI to AGI, will certainly require higher infrastructure hardware and large power consumption. Some researchers are starting to look at Blockchain as an alternative AGI acceleration solution. Decentralization makes all high infrastructure not have to gather in one place, besides that many web3 projects provide high AI compile support services. Such as RENDER, AKT, and others, one of which is Heurist Protocol, AI Cloud powered by Decentralized Compute. With decentralized compute services, it is possible to become an AGI acceleration platform.

AGI Support Infrastructure

Imagine in the future AI is able to learn, adapt, and make decisions based on the information received and develop it into unique information. To reach this phase requires more advanced AI evolution and large data covering the past. Processing this data requires computational power and deep AI infrastructure, computer GPUs also have a significant role in running AI from training to generating.

Decentralized AI-as-a-Service cloud, Heurist has become a project with high value for AI acceleration. Aggregated compute sources from ownership of GPUs and data centers to serve AI with serverless. Powerful GPUs are sometimes a bottleneck for AI developers in compiling complex AI infrastructure. Using API management for AI acceleration sometimes requires high costs and service limitations. With API-first, heurist removes dev bottlenecks in managing GPU machines, reduces costs, and censorship-free AI integration APIs.

Heurist in AI Development

Grand View Research predicts the Generative AI Chipset Market Size from 2024-2030 globally will increase by 32% CAGR driven by the growing interest of chipsets in the AI industry. The advancement of increasingly complex AI algorithms requires high computing power. Besides the AI industry, the robotics and autonomous vehicles industries also require similar chipsets, so it has the potential to exceed predictions by CAGR until 2030.

Heurist provides opportunities for limited chipset access for some developers. In general, Heurist products provide several offerings:

• Serverless AI-as-a-Service: Deploy and access AI with just a few lines of code connecting APIs and SDKs. In the computing process, sometimes the part of the workload that uses AI in it appears constraints on time-to-compute which affects the hardware power used, Heurist solves this problem by abstracting the complexity of hardware management.

• flexibility in compute resources and mining emission: As AI grows more complex, power efficiency on all sides is required. Heurist is able to manage the energy resource requirements of some tasks that require it, as well as manage the emission of tokens used in the network.

• Alignment-Centric Economics (ACE): The concept of economics expands computing resources, optimizes utilization, accelerates AI models, and prioritizes the long-term value of protocols. Heurist token utility is closely related to the entire interaction process in the ecosystem.

• AI Ecosystem in Web3: Heurist bridges AI into the crypto world. By supporting and incubating AI apps that are focused on crypto users both onchain and offchain apps.

Heurist Power Key

Become the “Airbnb of GPUs” where users can rent GPUs directly. Heurist is cutting through the technical obstacles of some people or groups regarding GPU power. However, there will be new obstacles in using Heurist cloud GPU resources.

Heurist offers the concept of serverless compute, with no need to install software or set up virtual machines. Adopting the Platform-as-a-Service (PaaS) model by abstracting the GPU power source dynamically according to user needs in real-time.

So that in general, all demands can be carried out by the Heurist platform. Not fixated on the use of AI and the development of AGI alone. As long as the demand has nothing to do with the connection between nodes in a network, Heurist is able to run power management based on the resources needed.

Unlock functionality as AI Image Generator

Heurist Imagine is a free-to-use app service provided by Heurist to create images with AI. The mechanism is the same as other generative image platforms, creating a prompt to generate an image as desired and supporting multiple finetuning of stable diffusion and flux models.

Find Anything in Internet Using AI

Heurist Search is a service that can be used to search for anything on the internet using the AI provided by Heurist. All things that you want to search using Heurist Search will be displayed in a summary form from several sources obtained on the internet and provide several questions or similar topics.

Your Helpful Assistant, Pondera

Pondera is a free service provided by Heurist and functions as an AI Assistant. As an assistant, Pondera can do many things with several selected AI models such as llama, qwen, mixtral, and hermes.

Heurist Protocol

As a decentralized GPU cloud platform, Heurist is built with a permissionless, efficient, and user-friendly system. Heurist protocol has several important elements such as Compute layer, Orchestration Layer, and Application Layer.

Compute Layer

The compute layer has several important elements that are the main foundation of the Heurist service. All elements are connected to each other with important functions in their parts.

1. Compute Nodes: Important infrastructure as the main foundation of GPU hardware and compute resources. There are several types of hardware that can be used as nodes, Single GPUs, Multi-GPUs Cluster, and Fractional GPUs (created from NVIDIA's Time Slicing and Multi-Instance GPU techniques). With it's permissionless feature, there is no limit to the node type specification. As long as it fits the minimum GPU specifications, making it a single GPU is allowed. This makes an ideal case for fair and balanced monetization of the hardware.

2. Pods: An important element as the brain that manages the compute resources that need to be used. Pods can manage hardware and software needs, so that needs can be adjusted to the resources owned and more efficient in using compute power. Pods are built from several elements, such as AI model inference, coordinating multiple AI models, fine-tuning-as-a-service software, and ZK proof generation service. This allows the compute layer protocol to organize active nodes based on needs and can streamline cost and resource allocation.

3. Validation System: Uses staking mechanism and several other methods as a form of crypto-economy. It's used to ensure the reliability of computational results and protocol integrity.

The native mechanism carried by Pods is auto scaling, the system can dynamically adjust the use of resources if it gets high demand and otherwise can reduce resource usage to 0 if it has no compute demand. The number of active nodes is set based on predefined metrics such as GPU utilization and the number of pending tasks, and is influenced by token voting. The dynamic mechanism of managing compute demand makes the processing requirements change based on the workload, making the protocol more efficient.

Orchestration Layer

Heurist is currently building a layer 2 chain to coordinate ecosystem and cloud operations, an elastic chain using the ZK stack. There are several benefits offered:

• High throughput but low costs: ZK technology is used to combine changes to the same storage slot over multiple transactions into a new data slot (state-diff model) aiming to reduce gas costs. A solution for micro-transactions that are common in cloud-computing services with pay-per-use pricing models.

• Interoperability: Aims for Heurist to interact with other chains without complicated bridge mechanisms. This will make it easier for Heurist to integrated services and partners.

• Sovereignty: The Heurist chain cannot be interrupted by outside parties, transactions that occur outside Heurist do not affect all activities within the Heurist ecosystem. This ensures reliable resource management, cloud operations, and payment systems.

A sequencer will be created at layer 2 which functions as a router that manages compute requirements based on hardware specifications, software compatibility, uptime, and efficiency. Transactions will be validated on the onchain to provide payments to related parties, when the task is successfully executed. If the request timeouts or is unsuccessful, the router is flexible to re-execute or discard the request with a failed execution remark, and no transaction is validated onchain before the task is successfully executed, reducing unnecessary gas fee usage.

Application Layer

Apps

Heurist is permissionless, anyone can build apps using the Heurist ecosystem. It doesn't have to be a decentralized app, it can be an app, API gateway, or integration on a dedicated business page. Heurist opens up opportunities to use its protocol in real-world use cases, by incentivizing developers and parties who contribute to the Heurist ecosystem. Most importantly, there is no central app that controls the entire Heurist ecosystem, in alignment with the concept of decentralization and sovereignty.

Payment Gateway

• For Developers: Adopts the payment mechanism commonly used by SaaS/PaaS application platforms, which is an upfront payment intended for compute credits. Credits are used to cover the cost of compute used, developers benefit from the spread between compute usage and service fees.

• For User: Users can easily use crypto wallet as a payment method. Payments are made instant based on the resources used, so there is no need to have a subscription to use the services of the Heurist ecosystem. This also allows new users such as AI Agents to pay for computing usage based on the resources used.

Utilize NVIDIA GPUs as Mining

Utilizing NVIDIA GPUs, we can earn HEU tokens by running AI Inference. The results obtained by miners are points designated for contribution and allocation in several phases.

🦙 Llama Points and 🧚‍♀️ Waifu Points

There are two categories of rewards, one is for contribution in image generation model and the other is Large Language Models (LLM).

• 🦙 Llama Points, designated for LLM miners

• 🧚‍♀️ Waifu Points, designated for Stable Diffusion miners.

Points earned for these two categories are based on the workload received by GPUs, hardware specifications, and latency.

Earn Waifu

Waifu points are earned from contributions in the generated image side. The proportion of point distribution is based on the pixels in the image and the number of iterations (sampling steps). There is a multipliers scheme based on the stable diffusion model used:

Tips about GPU series: Each consumer series (30-series and 40-series) has a compute power performance (measured using TFLOPs) of the GPU that keeps increasing as the last 2 digits of the GPU series are added. For example, 3060 < 3070 < 3080 < 3090; 4060 < 4070 < 4080 < 4090; which has a “Ti” series is slightly more powerful than the same series without “Ti”. While the Datacenter class is different; A4000 < A4500 < A5000 < A100 < H100.

Currently the NVIDIA GeForce RTX 4090 is the most available GPU on the Heurist network at 2908. High-performance GPUs such as NVIDIA H100 NVL, H100 80GB HBM3, and H100 PCIe are among the top 10 active devices on the Heurist network with more than 500 units, demonstrating Heurist credibility as a Decentralized AI-as-a-Service Cloud platform by providing a high level of architecture with a large number of units.

Earn Llama

Llama points are earned from the contribution of using the Large Language Model (LLM). To run the LLM model, VRAM capacity is very important in a GPU. The proportion of point sharing is based on the token length of the inference request. There is a multipliers scheme based on LLM model category.

Currently, there are about +14,000 active GPUs with a total VRAM of 511,608 GB and cumulative miners +22,000. This dashboard shows the sustainability of Heurist's network foundation, so that it can accommodate high workloads with consensus applied.

How To Participate?

1. Hold Your Own GPU: Previously explained about the GPU specifications that can be run in the mining section. The next step is to download the mining program at https://github.com/heurist-network/miner-release and set-up the mining node. For more details on specifications and step-by-step, please visit https://docs.heurist.ai/guides/miner-guide

2. Rent Managed Nodes: Realizing that not everyone has a GPU that fits the criteria, Heurist provides a special mining node service for the community with a more competitive price than the market price. Node service has been managed and organized by the team, so users can relax to get points to grow daily.

Heurist: Tokenomics

$HEU is Heurist Protocol' native token with a maximum supply of 1,000,000,000. The HEU token is used to trade all the values in Heurist Protocol, some of the utilities offered include:

• Paying for compute resources,

• Secure the network with a staking system and earn protocol fees,

• Voting power governance decisions, and

• Gas tokens for transactions in Heurist Chain

Heurist Protocol is still early, with a 10-year mining vesting schedule bringing Heurist the opportunity to improve the protocol mechanism in the future. Team & Advisory token distribution will also start unlocking in March 2025 slowly and linearly over 2 years until March 2027. There is an Untracked allocation of +6M HEU, these tokens will cliff 3 months and will be used as future grants.

Dynamic Emmision for Compute Providers

The allocation from emission protocol comes from 50% of the total HEU supply (excluding season 1 testnet mining), designed similar to Proof-of-Work (PoW) by ensuring compute resources are available and attracting new community members. Heurist as DePIN differs from PoW systems in that available resources must match the real demand for those resources. Heurist balances resource availability based on demand by controlling protocol emissions based on protocol revenue.

Before entering the mainnet phase, Heurist plans pre-mainnet mining rewards in 3 phases:

• Phase 1: 5% of the total supply is airdropped evenly to all miners and snapshot taken at July 19th, 2024.

• Phase 2: Using a weekly snapshot system, with a minimum amount of 400,000 and a maximum of 2,000,000 tokens per week, so that the tokens obtained by each miner will be dynamic depending on the demand from the GPU specifications used by the minner and several other criteria previously explained.

• Phase 3: TBD

The data collected in the 3 phases above will be used as a resource for consideration of the protocol emission mechanism during the mainnet. After the mainnet emission takes place, the team and governance have the right to determine the next emission schedule. The emission rate is determined from 2 parameters max emission rate 5% per year and baseline emission rate is 1.25% per year. If at any time there is a significant decrease in demand, emissions will naturally decrease as well to prevent oversupply of resources. If there is a high increase in demand, emissions will be increased to provide more resources to solve the high demand. This dynamic adjustment aims to create a flexible and responsive protocol to incentivize network participation.

Heurist: Accelerate Agent Onchain

Heurist launched the Agent Framework https://agent.heurist.ai/, using a lightweight framework written in Python. The Agent Framework is integrated with Heurist's AI-as-a-Service (AIaaS), connected to decentralized protocols, giving builders the tools to build AI Agents with various utilities such as trading assets, market analysis, and communicating with the community.

Unlock “The Raspberry Pi” of Agent Frameworks

If Eliza is the Ubuntu of open source agent frameworks, Heurist Agent Framework is the Raspberry Pi OS with a lightweight, fast, functional, cost-effective framework to scale decentralized AI. Heurist framework architecture is less complex, solves the main obstacles in developing Ai agents, and does not have a multi-sheet framework guide. Built using the Python programming language, so it is compatible with the AI ecosystem, the framework is also lightweight and suitable for beginners but powerful enough for production deployment. Although the Heurist framework seems simple, there are no limitations in future development using the Heurist Framework.

Plug and Play

Heurist framework is designed modularly to suit the needs in development, the basic framework is quite lightweight and can be further developed using the given framework. Built using the Python programming language, where 80% of AI tools and algorithms use Python and this makes the framework suitable for the AI/ML ecosystem.

Ease Prompt Engineering

Agents cannot look like just bots, prompt engineering is required as the key for agents to operate autonomously and decentralized. The Heurist framework makes it easy to edit prompts, simulate the responses, and track intermediate states for debugging.

Inference Scalling

The term “Inference Scaling” can consist of several techniques such as chain-of-thought reasoning, ReAct prompting and self-consistency methods (same like OpenAI-o1). This algorithm natively supports more intelligent agents than bots. The use of high inference causes computational overhead, but can be minimized by cost reduction from open-source LLM and decentralized compute resources.

Multi-Platform Support

Agents can communicate on multi-platforms, commonly agents can interact on Twitter, Telegram, and Discord. Heurist provides more value with agent interaction in Farcaster and Lens Protocol with just one deployment.

Infrastructure Partner

There are several important partners in supporting the Heurist accelerate Agent Framework such as HoloWorld for creating 3D and AR/VR, Akave for storing and monetizing data using Filecoin, Gateway Network for executing private computations in TEE, GenLayer for deploying and interacting with AI-powered smart contracts, Spheron Network for accessing raw GPU compute, Space and Time for accessing ZK-proven trusted data sources, and ZKnync for transact seamlessly in the Elastic Chain ecosystem.

Heurist will go as far as?

Encouraging the development of AI that is advancing towards AGI technology requires a lot of resources. Heurist Protocol is one of the providers of decentralized GPUs which are important resources in the development of AGI.

With a potential CAGR of 32% by 2030 in the generative ai chipset market cap sector, Heurist has a great opportunity to become a provider of Cloud GPUs. Heurist Protokol can jump into several narratives, DePIN as a cloud GPUs hardware provider that is managed in a decentralized and elastic chain mechanism that can manage resource usage emissions based on demand and protocol revenue. Can also dive into the narrative of AI and DeAI as a cloud compute resources service provider, not many builders have powerful hardware to execute increasingly complex AI programs. Heurist can also jump into the AI Agent narrative with a lighter, faster, functional, cost-effective Heurist Framework for decentralized AI scale.

I believe Heurist Protocol can find its momentum as a decentralized cloud GPUs provider. Based on the potential Heurist will be needed in the future and the strength of the Heurist chain infrastructure is proven by the total VRAM, active GPU, and the list of top 10 active GPUs, some of which are high-end NVIDIA GPUs.

As a newcomer in the GPU provider, Heurist will certainly get some challenges. Seeing from some big competitors such as RENDER, Akash Network (AKT), FLUX, Golem (GLM), iExec (RLC), and others who have established earlier. However, Heurist can be an interesting protocol in Agent AI.

Messari released an overview of Agent Frameworks, Heurist is quite interesting with its Edge in full-stack DeAI. AI Agent, which is currently still an online influencer, certainly needs utility expansion that is more useful for many people. With increasingly complex utilities, Heurist can provide the compute resources needed in making complex AI Agents.

Proven by the ai16z x Heurist partnership, The Decentralized AI journey. Heurist plays a role in providing serverless AI integration - zero GPU/CUDA headches, Free API credits, technical, and BD support, and academic research collaboration. This partnership can be a synergy for Heurist in supporting the growing Agent ecosystem with complex utility.

“This article is for education and personal opinion only, not a suggestion to buy. Please use this article as your DYOR resource for self-actualization or decision-making resource reference.”

• Transforming Research with DeSci

• Funding and Collaboration Solutions

• Innovation Through Tokenization

• Future Potential of DeSci

  ---

DeSci (Decentralized Science) started to gain mindshare during Q4 2024, after a video of CZ and several founders getting together to discuss the DeSci opportunity. The DeSci narrative started to rise following that moment.

In this context, I want to talk about DeSci potential and future development opportunities. Focusing only on ResearchHub, a project that I think has high chances of future development. whether in terms of protocols, teams, mechanisms, or novelty.

DeSci exists to create a new scientific research infrastructure that is public, transparent, and accessible to everyone. Utilizing blockchain technology to leverage infrastructure and create new methods in the scientific research process. This solution comes from the challenges and limitations of Traditional Science Research (TradSci), among others:

Centralized Funding: Funding plays an important role in the entire research process. Scientists spend more time applying for funding than doing research. Generally, funding is obtained from centralized institutions or grant agencies. This leads to a tendency for funders to interfere with decision-making, research processes, intellectual property, and resource allocation.

Difficult to Access: Many scholarly works are not freely accessible, locked in the form of subscriptions or paywalls. Some of them are not affordable for individual researchers. Limited access makes it difficult for researchers to stay up-to-date and keep up-to-date with information or discoveries.

Lack of Real-World Impact: The research is focused on preserving the status quo rather than challenging it. Some researchers use the same tools and methods over and over again, only varying the samples to increase positive reviews. The academic system emphasizes quantity over quality. This is why research results generally do not have an impact on the world.

Blockchain as Leverages

DeSci stands in the same place as DeAI (Decentralized AI). Utilizing blockchain technology to leverage the functionality already done in web2. DeSci is able to answer TradSci common analogy “Valley of Death”, which refers to the phase of scientific research results that have no impact on industry and academic novelty.

The “Valley of Death” phenomenon makes scientific research results have no clear impact on the world. Researchers only focus on results that attract attention without reviewing the impact provided, just enough to increase the chances of being accepted in journals. The results of scientific research that does not make sense, so it raises doubts about its accuracy.

So, how does DeSci solve these challenges?

Decentralized Funding; Provides a solution to the problem of funding research. The community deserves the right to choose research proposals that have merit. This mechanism also saves the researcher from having to spend a lot of time looking for funding. The benefit of decentralized funding is that it creates a research focus that has an impact on the world, whether for knowledge or industry.

Decentralized Research; ResearchHub addresses the challenge of the distribution of research that cannot be accessed for free. Its public nature makes it easy for everyone to access for actualization. This hub makes it easier for members to spread awareness, value research, and mindshare.

Token Incentive; An option that can be taken to initialize research into a token. This token can be used as an incentive method to members who provide support to research projects. This incentive model can be further improved to increase mindshare and sustainability of research in the future.

DAO; Community members can contribute to a research project even though they are not part of the core team. Conduct peer reviews, vote on research proposals based on power, and others. The general DAO mechanism in web3 is suitable for use in this narrative. Knowledge for everyone.

Research Hub: Perfect Example of DeSci Fundamental.

ResearchHub is a platform to accelerate scientific research. A platform for communities that understand the world's needs and problems in traditional research. The funding mechanism is the main concern why this platform exists, the lack of objective metrics in research, and the profit business model of academic publishers.

ResearchCoin (RSC) Follows Quality Scientific Research

Build on the Ethereum ERC20 token, used for digital assets of the scientific community. If you understand how peer-review works in academic publishing. You will certainly agree that peer-review is broken with a private and centralized system. ResearchHub highly values the science produced by researchers. RSC as a digital asset is used to reward researchers for their contributions to science - regardless of whether their research fits into the existing structure.

RSC tokens are used to fund experiments and collaborations. Earning RSC is easy by contributing to the ResearchHub platform. Other ways are by sharing research papers, discussion, peer-reviewing, receiving RSC from others, and developing open-source features.

Tokenomics

Implementing the Web3 Sustainability Loop, the tokenomics RSC is built for long-term self-sustainability. ResearchHub has a lot to do with grants, incentives, and rewards. It needs a special mechanism to keep the ecosystem surviving.

The total supply of RSC is 1,000,000,000 and will be distributed all in the range of 12 years with the distribution as shown above. The community gets a 60% portion where the maximum emission distribution rate is 5% per year. This is to adjust the incentives received by users and contributors with the platform's long term goals.

From the “Genesis Supply” vesting schedule is applied to the foundation, founders, new employees, and as needed to the Hot Wallet. The hot wallet is used to pay platform editors and users. Genesis supply also quarterly provides allocation to Treasury. Treasury is used for Uniswap LP assets and for monthly foundation allocations based on RIP proposal approval. Treasury gets inflow from Grants and user bounty fees.

Thesis

The DeSci narrative has been initialized in 2020, starting to gain mindshare since the AI Agent launchpad and Meme launchpad narratives were busy coupled with a meeting of several founders discussing DeSci at BUILDer House Bangkok. But long before that, Brian Armstrong and CZ had already stated the potential of blockchain integration in science acceleration systems.

Momentum of DeSci is getting clearer with the initiation of pumpdotscience as a research experiment platform and several DeSci-related ai agents. Some of my personal thesis related to DeSci:

• I believe this narrative is not for everyone, there will be many new groups that have the same vision and mission as the fundamentals of DeSci. Building a DeSci ecosystem that is welcoming to all new scientific ideas is difficult to do in traditional science.

• Memes will become the gateway to the DeSci narrative, plus the utility of DeSci's AI Agent. Flowing liquidity and AI Agent help and facilitate the research process. Launchpad and DAO DeSci will increasingly attract the attention of many researchers.

• Opening up space for massive collaboration, the research process will be more diverse, resulting in new knowledge that comes from many thoughts. New knowledge will emerge that can be created from diverse collaborations.

• Will leverages DAO and Smart Contract used in decentralized funding. Allows for IP tokenization (in the form of NFTs) nor research coins for research development. Decentralized funding means public access, the results of experiments or research can be accessed by the public or community members. Allows for faster and less costly knowledge transmission. Transparency of the research process includes data usage, allocation of funds, research logs, and peer reviews.

Conclusion

Fundamental DeSci solves TradSci problems that stem from funding. By using blockchain as leverage, the DeSci platform can develop a more fair, transparent, and decentralized solution mechanism. Of course, at the moment, much of the DeSci platform is still raw and experimental, but it has great potential in the future.

I believe ResearchHub can be a pioneer in creating a platform that solves the problems of TradSci. Decentralized funding, transparency of research results, DAO, and incentive models in one platform are new things in academic research. It's very possible that there will be new science that stems from the DeSci experiment. Because the freedom and needs that many people want support a massive movement in science.

“This article is for education and personal opinion only, not a suggestion to buy. Please use this article as your DYOR resource for self-actualization or decision-making resource reference.”

Smart Contract but with LLM

GenLayer builds a smart contract mechanism with an AI-powered smart contract design called Intelligent Contact. Intelligent Contact can interpret and execute instructions written in natural language. This mechanism is capable of performing more complex operations beyond traditional smart contracts. The Large Language Model (LLM) used opens the gateway to access and use real time web data, creating dynamism in response to external events and data sources. GenLayer uses Optimistic Democracy as a consensus mechanism. This mechanism makes it possible to strengthen security and efficiency by appealing and re-evaluating if something goes wrong.

Intelligent Contracts

Intelligent Contracts are ai-powered smart contracts that use LLM to access real time web data and understand natural language. The difference from other smart contracts is the dynamic nature of interacting with external data and adapting to changing data conditions. It can then modify its operations as new information becomes available.

Intelligent Contracts also open up new use cases in a wider scope than traditional smart contracts. Intelligent contracts support dynamic applications like smart oracles, interactive games, and AI-driven decentralized applications. In addition, intelligent contracts use the Python programming language which is more familiar to builders.

Optimistic Democracy Consensus Mechanism

This consensus applies an equality mechanism and a structured appeal process. Its function is to reach consensus on transactions that are subjective and non-deterministic. Here is an illustration of how consensus works:

When a transaction is sent to the consensus mechanism. A random validator group of 5 validators is created to verify the transaction. The number of validator groups can be adjusted to increase security in consensus.

Randomly, one validator in the group will be taken as the leader. The leader will execute the transaction and submit an output. On the other hand, other validators in the grub test the leader's output using the equivalence principle. If most validators accept the output proposed by the leader, then the transaction is provisionally considered accepted. However, the decision is not set in stone, it is possible to appeal for a limited period of time, known as the Finality Window. If a vote is not reached, the subnet grub will be replaced by a new grub and repeat the process.

The initiating and appealing process occurs during the finality window. A new validator grub is formed with the task of testing the proposed result that was deemed acceptable earlier. This new grub will vote on the proposed result, if approved then the transaction is accepted, if not then a new subnet will be formed with 2x the number of the previous subnet to increase accuracy.

The appeal process will continue until a majority consensus approves the transaction. The decision will be recorded and the transaction status updated. Only validators who are recorded as successfully completing the transaction will be rewarded.

Equivalence Principle Mechanism

Equivalence Principle is a mechanism that allows intelligent contracts to work consistently across all validators when handling non-deterministic outputs. Non-deterministic operations allow the intelligent contract to interact with the LLM and access web data. This mechanism is critical to the way validators validate and vote on the results proposed by the leader.

The figure above is a schematic example of the comparative equivalence principle. Essentially, validators work to reach a consensus on whether the leader's results are acceptable, this decision may involve direct comparison or qualitative evaluation, depending on the contract. This principle is fundamental to intelligent contracts, ensuring the contract can work with different validator networks.

• Consistency in Decentralized Output: The Equivalence Principle makes it possible to display output from a wide variety of sources such as LLM or web data. Dynamic data processing mechanisms are still considered valid as long as they meet the standards set in the contract. This is important for maintaining fairness and uniform decision-making on various blockchains.

• Security Enhancement: The Equivalence Principle requires each validator to validate the output of another validator. This two-way verification helps prevent errors and manipulation, and ensures data is accurate and agreed upon based on the contract that has been created.

• Output Validation Flexibility; Intelligence Contracts will inevitably handle complex data. This principle allows the builder to create rules about what kind of output is “equivalent” or acceptable. Ultimately, each ecosystem in the GenLayer does not have the same contact specifications, due to different needs.

Conclusion

GenLayer offers a new solution in smart contract mechanism. AI-powered intelligent contracts allow smart contract mechanisms to directly interact with web data. In addition, GenLayer Optimistic Democracy Consensus uses the Equivalent Principle to execute transactions that are subjective and non-deterministic. This consensus makes the time-per-finality (TPF) of intelligent contracts longer to accommodate appeals and improve reliability. Unlike Ethereum, where finalization is achieved in 6-12 minutes on average, GenLayer finalization is achieved in 30 minutes on average.

“This article is for education and personal opinion only, not a suggestion to buy. Please use this article as your DYOR resource for self-actualization or decision-making resource reference.”

• Virtuals Fun as a solution to the problem of high IAO slippage

• Revenue generation system from the Agent side, User, and Contributor

• Opinion about bottlenecks with liquidity and tech

LUNA is the first proof of Virtuals success in creating a good launchpad for AI Agent. The rise of LUNA is also in the right momentum, previously the AI narrative was on the rise. Several AI tokens have received massive liquidity. But the liquidity to memecoin AI has not yet gotten a good position. Virtuals Protocol comes as pump.fun for AI Agents, designed as a place for owner freedom in creating AI Agents with systematics similar to pump fun.

Virtuals Fun emerged to solve the IAO (Initial Agent Offering) problem, initially there was a problem when listing Agent tokens. Early users who want to buy the token have problems with high slippage. Virtuals Fun also makes it easy to create ai agents without an ai background.

I see that the AI Agent narrative will continue to run until the end of this bullrun. I believe Virtuals is able to continue to develop following the roadmap, narrative, and whitepaper. Virtuals background supports the existence of AI Agent needs to continue to exist and develop greatly in the future.

Virtuals: Revenue Generation

Virtuals is building an ai agent that focuses on gaming and entertainment using the concept of co-ownership. The technology brings unique capabilities such as acting on assigned tasks, being able to communicate using voice or text, and being able to be integrated on other platforms.

Agent Usage Fees are predetermined in the per-inference cost process. Everyone can use public API Access to use the agent permissionlessly. The payment mechanism requires the user to have $VIRTUAL tokens in his wallet and the number of tokens will decrease as the agent is used on-chain.

Revenue flow is done by the user when calling the agent via the API, $VIRTUAL tokens will be deposited into the agent wallet. Then the collected $VIRTUAL tokens will be used to buy back agent tokens from the open market. Then the purchased agent tokens will be burned. Agent tokens that are burned will reduce supply and create deflationary pressure. This mechanism will increase the value of agent tokens. While the IP owner of the agent will receive revenue from trading fees liquidity pools of agent tokens.

The diagram above explains the concept of how co-contributions can help improve AI Agents using the Modular Consensus Framework. This framework is also part of the Virtuals Protocol architecture. Its purpose is to facilitate the construction, maintenance, governance, hosting, and utilization of protocol-managed VIRTUAL Agents. Its nature is modular to pursue transparency and compatibility. Each contribution from co-contributors that has been validated, will get an NFT stored in the Immutable Contribution Vault. These NFTs become proof of collaborative efforts and intellectual contributions in the ecosystem.

Virtuals: Bottlenecks with Liquidity and Tech

• Massive Ease of Future Rugged Dev

At a time when the AI agent narrative is on the rise, there comes a lot of liquidity from many directions. For new agent tokens in Virtuals, slippage becomes a problem from the low token cap that comes with a massive number of users. This problem has been resolved by the Virtuals Fun method at the IAO stage. However, the introduction brought by Virtuals related to Virtuals Fun which is similar to pump.fun creates bias. Bias in the ease of making agents meet platforms that are not necessarily in accordance with the vision brought by Virtuals Protocol.

Of course this is good for increasing the adoption of AI Agents. But the conditions will be different when the agent has been deposited into Uniswap and burned (sentient agents). The ease of creating agents at the beginning becomes a double-edged knife for the owner to develop the agent. The difference in vision makes the dev or owner give up or leave the agent intentionally (Attention + Meme + AI Agent + Owner Low Experience = Rugged, imo). In the end it will only be PvP AI Agent narrative.

• Revenue Generation Limitations for Non-Gaming/Entertainment Agents

Virtuals announced to play in the gaming and entertainment sectors. Bringing a unique approach and belief in my opinion, “Dopamine Loop” and “Capitalism as a Decentralizing Force”. However, with the concept of freedom in creating agents, Virtuals platforms are now not only filled with agents that are linear with the concept.

At this stage, sentient agents that are not linear with the Virtuals concept are still not able to unlock maximum revenue generation. Of course Virtuals will announce ecosystem partners that will support some of these sentient agents. Until waiting for an ecosystem that can be used by agents outside of vision, the revenue generation system has not worked optimally.

• The Connection of Each Sentient Agent

Each agent has its own functionality, but it would be good if there was a mechanism that could connect each agent to be able to interact. Quite complex but interconnected agents can form a mechanism of agents that move together. So that the ecosystem becomes more solid and mutually beneficial.

Conclusion

Virtuals is creating momentum amidst the rising AI narrative. Significant innovation in AI Agent development through a platform built with ease of agent creation with a structure that resembles pump fun. Mechanisms using native Virtuals tokens as inference costs and revenue flow based funding can strengthen this ecosystem with deflationary pressure through the agent token burning process. There are some technical and liquidity challenges that I think could arise in the future, such as ease of Rugged Dev, Limited Revenue Generation for Non-Gaming/Entertainment Agents, and lack of Inter-Agent interconnection.

Must Read:

https://whitepaper.virtuals.io/

https://x.com/0xMrDiaz/status/1849027402774827125

https://4pillars.io/en/opinions/virtualsprotocol

https://4pillars.io/en/opinions/virtualsfun

“This article is for education and personal opinion only, not a suggestion to buy. Please use this article as your DYOR resource for self-actualization or decision-making resource reference.”

The Open Network (TON) is one of the Layer 1 platforms that has attracted tremendous interest in the crypto industry. The large increase in TON is related to it’s integration with Telegram. Telegram opens up space for platform development and integration of Web 3 technologies, this value makes Telegram different from other social media platforms.

Data from Artemis shows that the daily activity address on the TON Network since the beginning of the year has increased significantly. TON got the right moment in 2024, P2E on Telegram brought massive new users. The highest peak of user activity on TON occurred on September 30, 2024 with total of 1.4m users. In addition, the integration of Tether USDT which supports the TON Network in April 2024 also brings convenience to Stablecoin access.

Currently there are more than 1300 dApps in the TON Ecosystem with various categories. Many of these dApps are participants of The Open League program. The Open League was originally a Telegram project that was later opened to the open community. TON provides a space for various decentralized applications and services. The Open League operates like competitions, leagues, and other campaigns with the aim of raising awareness of blockchain. In March 2024, 30 million TON was awarded as community incentive ($115 million at the time)

• $38m for miners and user incentives

• $22m for airdrops

• $155m to support League Developer Ecosystem

• $40m boosting liquidity pools

History

Looking back at the ICO boom, the largest ICO was conducted by Block.one the creator of the EOS Network, raising $4b in 2018. Telegram’s ICO was the second largest, with $1.7b raised mostly privately from a significant number of private investors. The funds raised will be used to develop Telegram’s blockchain ecosystem, in line with Telegram’s mission to create a decentralized platform that goes beyond messaging.

In 2019, The U.S. Securities and Exchange Commission (SEC) filed a lawsuit against Telegram, regarding the ICO conducted as an unregistered securities offering. This condition made Telegram to temporarily stop the ongoing project and return investors in 2020. The cessation of the project doesn’t stop the mission from the initial concept that has been made.

TON Blockchain has been restarted with a $6m investment round led by Rune Capital and RTP Global. Funding will be used for the development of TON Blockchain. RTP executive, Stepen Ivanov said,

“We believe TON Labs can finally take the blockchain from the niche and geeky space today to a mainstream technology that strongly complements the current IT landscape”

Furthermore, several funding until 2024 was done by several tier 2 and 1 VCs such as Pantera Capital, Animoca Brands, DWF Labs, etc.

Transaction activity on the TON Network began to increase very significantly in the December 12-14, 2023 date range. The increase in on-chain activity occurred because at the time there were several events that occurred in TON. Previously, maintenance occurred due to exponential growth in TON Network activity, then there were cNFT events, Ledger hard wallet support TON, and TONomania Phase 1 for CIS users.

In addition, P2E projects on TON also have a huge impact on the sentiment of TON Network. After NOT Coin listing, many retailers speculated that all P2E projects on TON Network were “probably something”. The number of users who join the P2E campaign, in addition to benefiting network growth, also provides benefits to Telegram as a platform.

Revenue and daily user activity on TON Network increased following the news that TON became ads payments exclusive on Telegram. This news also created a new bullish narrative and wave on the TON Network. Hype brought by P2E also has an impact on other projects that make TON a ‘home’ such as DeFi projects, NFTs, DAOs, etc.

TON as Layer 1

TON is a Layer 1 blockchain with Proof of Stake (PoS) Consensus. Unlike a monolithic network like ETH, TON consist of a MasterChain, which includes several WorkChain, and a ShardChain that aims to create blocks ‘vertically’ to achieve efficiency.

TON uses Byzatine Fault Tolerant (BFT) consensus mechanism ‘Catchain Consensus’ capable of improving security and reliability. This consensus makes it possible to complete transactions even if other nodes act maliciously of fail to respond. The combination of BFT and Pos creates a strong framework for maintaining network integrity and preventing attacks.

The Sharding Technology used is with a dynamic sharding mechanism, making the blockchain able to split and merge based on load so that transactions and smart contracts are more efficient. TON has an Infinite Sharding Paradigm that makes the blockchain able to accommodate many transactions without sacrificing network performance. One of the advantages of TON is Instant Hypercube Routing which makes transaction spread across the network to minimize latency and improve UX.

Validators who wish to participate on the TON Network must stake a minimum of 300,000 TON, the rewards given are based on block rewards and transaction fees. While penalties include confiscation of some or all of the staked TON for signing invalid transactions. Some of the confiscated funds will be burned, while some will be given to the Fisherman who reported the invalid block.

Behind it are Nominators who can delegate coins to validators, sharing the risk and revenue associated with validation. Fisherman can be any node on the network, can store TON on the MasterChain and report invalid blocks. When validators misbehave, some funds are confiscated and some are sent to the Fisherman and burned. Collators are responsible for suggesting new blocks to validators, a task usually performed by the validators themselves.

TON On-Chain Data

Based on token terminal data, TON is currently in 6th place as an L1 blockchain project based on fees since the beginning of the year. It was good year for TON which was able to compete with other L1s that have already run their utilities. As of October 28, 2024, TON collected a total of $29.496m in fees YTD.

As a business, TON performed quite well in Q3 2024. Revenue of $6m and Fees if $12m increased by 14% compared to Q2 2024. Token holders also saw an increase of 147% or 48m compared to Q2. KPIs that include user activity, core developers, and code commits have also increased compared Q2, this shows that TON both in the core team and user activity are still working hard to improve the performance of the TON Network.

The fees earned by TON have increased several times, the increase in fees certainly goes hand in hand in hand with the increase in on-chain activity. The average fees earned by TON YTD are around $97,994 USD per day. It’s interesting to know that if on-chain activity increase, fees can increase up to 2x from the average.

On-chain wallet activation shows smart-contract wallets that make transactions and active wallets show wallets that make transactions. Both seem to have the same increase, until there is a different on October 20th. On-chain wallets experienced an increase in activity, but active wallets didn’t have a significant increase like on-chain wallets.

Cumulatively, TON TVL since August has decreased very significantly. Indicating a decrease in user confidence to stake into the TON network. In addition, Mcap Stablecoin has continued to increase to 860m USD since October 18. The increase in Stablecoin Mcap shows TON support for Stablecoin access into the TON network.

Based on price if TON gets good momentum, it’s easy to reach the previous ATH at $6. All price related possibilities have may POVs. But i’m personally optimistic that TON will be able to touch $6 again.

Self Opinion

• TON still has great potential, many events and incentive campaigns that benefit builders and retails. I believe this can be an attractive mini app catalyst to bring in many new users.

• Telegram as messaging platform is still in high demand by users, it’s likely that TON adoption will increase in line with Telegram’s growth. In addition, Pantera Capital mentioned that TON is their biggest investment ever during it’s establishment. This condition makes TON easy to lead further with Pantera who always participates in TON development activities and progress.

  

• TON still focusing on it’s own platform and infrastructure, it’s still difficult to use TON on other networks. This is a challenge to bring web 2 to web 3 and inviting native crypto into the TON Network.

• Telegram can be the best place for Games, DAOs, and NFTs. I personally don’t see Telegram being a good platform for DeFi based on the mechanism of Telegram as messenger app and the functionality that Telegram brings.

Conclusions

Like the process of computerized statistical theory, TON x Telegram solves a new method of applying crypto mechanisms. With high user activity on Telegram, it provides easy exposure for TON to grow through new users. TON x Telegram has a profound impact on the way the project introduces the developed utility. The platform is easy and widely used by many people, and the functions brought by Telegram are in line with the concept of bridging web 2 users to web 3.

TON's ambition to become a “SuperApp” is clear. It’s very possible for this to happen on the Telegram platform plus the support of some top tier VCs. Maybe what it looks like today will not be the same as 2025. With the rapid movement of the crypto space, it’s interesting to see how TON adapts to the conditions and adjusts to the roadmap that has been made.

Currently TON as an ecosystem is still in the early stages, many things have not been realized as a “SuperApp”. The Open League may be a leverage for TON to develop a more ecosystem. However, a big criticism related to The Open League is the different visions of projects that are members of TON as an ecosystem. Many projects that work with user exposure tend to try to “enrich themselves”. The sentiment of crypto native users can certainly have a negative impact on TON.

It’s difficult to project the potential of layer 1 projects that are still in the early stage. Only time and the performance of the TON core team will tell in the end. TON x Telegram as a web 3 mechanism concept already provides a massive change in the way the project promotes its utility.TON brings new opportunities, new mechanisms, and new infrastructure that are useful for blockchain development.

“This article is for education and personal opinion only, not a suggestion to buy. Please use this article as your DYOR resource for self-actualization or decision-making resource reference.”

References:

• https://medium.com/@tongems/airdrops-the-open-league-ton-and-how-to-get-ton-a05efb917633

• https://www.sec.gov/newsroom/press-releases/2020-146

• https://cointelegraph.com/news/ethereum-ico-boom-history-crypto

• https://cointelegraph.com/news/ton-labs-raises-6m-in-support-of-free-ton-blockchain

• https://icoanalytics.org/projects/the-open-network/

• https://panteracapital.com/blockchain-letter/ton-our-largest-investment-ever/

• https://crypto.com/university/ton-coin-telegram-how-to-buy-ton

• https://cointelegraph.com/news/breaking-telegram-abandons-telegram-open-network-and-gram-tokens

• https://en.cryptonomist.ch/2023/11/12/telegram-open-network-heres-how-works/

• https://medium.com/coinmonks/ton-tokenomics-an-overview-1d63413b1ebc

• https://4pillars.io/en/articles/ton-too-big-to-ignore

Layer3 will be expanding to Solana in October, with a new platform, migration token options, and a better UX. They are trying to create a new “Consumer Hub” in Solana along with 25 other top protocols. The goals is simple, provide a Solana user platform for L3 products, create a consumer distribution layer for all Solana protocols, and provide keys for Solana devs to play in the L3 ecosystem. This expansion will use Wormhole utility functions to support cross-chain users.

Attention and Identity for a User-Owned Future

Layer3 is a “Quest Platform”, the infrastructure offered as a Consumer Hub is currently irrelevant to the advancement of blockchain technology today in my personal view. But now Layer3 is bringing new changes to its system by building a user-focused infrastructure, supported by Omnichain identity and distributed protocols. Build a network of identity, incentives, and interface networks owned by the community. Layer3 must maintain a good UX to make it easier for users to interact cross-chain, as a Consumer Hub platform.

Recently, many new projects have emerged whose utility has not been widely used “including decentralized identity”, utilizing the attention of retail to maintain the stability of the project's income. Plus the contribution of AI in the crypto space, Layer3 believes that AI can exacerbate this problem. Related to capturing user attention, it’s necessary to make some more critical strategies.

Layer 3 is trying to stand strong by trying to create:

• Projects to reach out & build communities and provide value to users who contribute to the development of the project,

• Users to explore the protocol and earn from perpetual incentives, pre-token networks, and cross-chain identities.

The new Layer3 infrastructure has 3 components consisting of Identity, Incentives, and Interfaces.

All components are decentralized and owned by users in Layer3. Ownership and control are given to the community, so Layer3 can open up the ecosystem for users to get value from what they create.

• Identity; CUBEs are the fastest-growing multichain identity network, accelerated by AI. CUBEs serve as a connected onchain identity by uniting identities and activities across the chain,

• Incentive; reward users using points or tokens directly on Layer3. Each point or airdrop program created becomes a market value for Layer3,

• Interface; Layer3 as a consumer hub, has an easy interface for retail users and protocols that support it.

Layer3 catalyst as a platform for users with decentralized identities, incentives, and interfaces empowers users and projects to grow and create mutually beneficial value. Layer3 provides the infrastructure for users to participate or projects to build communities.

CUBEs

Layer3 introduces CUBEs as a decentralized identity based on cross-chain identity system. This system helps users to use one identity for various blockchain networks. The CUBE system is similar to a digital passport, users can move between ecosystems without the need for identity verification again. Decentralized identity gives users control over identity data in their hands without using a centralized third party.

Layer3 supports the idea of a user-owned KYC credential system, which enhances privacy and decentralization. Users can control their credentials and projects can interact with verified identities without the help of a centralized party. This decentralized identity infrastructure is directly linked to the Layer3 ecosystem which allows users to earn rewards for maintaining their identities across multiple blockchain platforms.

There are more than 6.5M unique Users, 95% of which use EVM wallets. Almost 2M unique credential holders and 29M cube mints. Layer3 onchain overview shows the amount of user participation in the infrastructure built by Layer3.

WorldCoin (WLD): Fair Competitor

WorldCoin makes technology more specific in how it verifies identity. WorldID (Orb), a decentralized identity verification system that uses biometric data collected using a special tool, the “Orb” obtained from iris scans. It brings a proof-of-personhood system for human verification and uniqueness. WorldID preserves the uniqueness of each identity from identity duplication using a decentralized ID system, a solution for future AI developments.

Key Compete

• Technology; WLD uses biometric data for proof-of-personhood, while Layer3 uses decentralized credentials for multichain activities with the same identity,

• Verification Method; WLD requires physical verification “iris scans”, Layer3 only digital verification. Both have drawbacks such as WLD dealing with the use of personal data, Layer3 with identity fragmentation issues and leveraging trust in on-chain data,

• Privacy; WLD triggers concerns because it uses personal data, namely irises, which is usually done by basic web2 government systems. Layer3 brings privacy security by not using personal biometric information.

Tokenomics & Utility

L3 goes live in July 2024 with 100% community ownership. Fair launch with total ownership by users and further unluck next year.

• Total Supply: 3,333,333,333 tokens

• Contract: 0x88909D489678dD17aA6D9609F89B0419Bf78FD9a

• Network: Ethereum mainnet

The distribution is done mostly for the community. 51% or 1.69 billion L3, core contributors 25.3% (843 million L3), investors 23.2% (773 million L3), and advisors 0.5% (16 billion L3).

L3 foundation has the right to buy back L3 using protocol income to maximize community benefits. In addition, there is a burn / stake mechanism with the following mechanism:

• Community votes to initiate revenue streams to stakers or buy & burn mechanism revenue protocol,

• Burning or staking L3 to access posting quests, deploying incentives, and accessing the CUBE credential network.

L3 has an interesting staking feature, known as Layered Staking. Where users can maximize revenue by staking and actively participating in Layer3. By staking in L3, users can earn passive yield into the staked amount. The Layered Staking mechanism opens up opportunities for users to earn more L3, fostering long-term engagement and community growth.

Layer 1: Passive Staking Rewards & Governance

Users earn passive earnings by staking L3 tokens to earn additional L3. Restaking concept but in the same token as the reward. L3 tokens can also be used to participate in governance decisions.

Layer 2: Actively Earn Other Tokens and Protocol Utility

Use L3 tokens to participate in exclusive stake-to-earn quests. Increase levels by increasing the number of staked L3 tokens. Get new early access projects on Launchpad access that are designed as pre-airdrop campaigns. Participate in other incentive programs that certain tiers can participate in based on the number of staked tokens.

Layer 3: Actively Earn L3

Users can increase the airdrop portion if they are more active than other users. For example, if a user completes 10 quests, they will get a 1.5x multiplier on L3 rewards, if they complete 20 quests, they will get a 2x multiplier.

Layer3: Personal Thought

Comparing Layer3 with other projects with the same narrative also doesn't make the comparison fair because the majority of projects in decentralized identity move in their respective areas. Proof-of-humanity is too broad a segment to say that project X is a pioneer, project X is an innovation, etc.

In my personal view, here are some of the things that make Layer3 important:

• Consumer Hub; Layer3 is able to open access for all new projects from all chains to Layer3 users. Decentralized Identity (DID) mechanism makes Layer3 all-in-one project and easy cross-platform interactions.

• Decentralized Identity; the use of decentralized Identity (DID) in 25 protocols certainly provides convenience in the multichain identity process. But it is very likely to occur problems such as protocol security, privacy, and unmonitored transaction fees.

• Multi-Earn Rewards; Layer3 prioritizes users with a large allocation to the community. Layered Staking mechanisms also favor users with multiplier rewards, projects and users benefit from each other in this situation.

The Decentralized Identity narrative will be very necessary in the future seeing the massive development of AI. Mitigating the risk of such a large AI impact can be done with the DID system. I'm sure there will be a time when the DID narrative becomes very large, but for now I'm not so sure its use is massive and in a general context not only crypto space.

always look up the development of DID projects that bring changes in the future. The adoption, integration and partnership of DID projects in the future will be potential along with the development of blockchain technology.

“This article is only for education and personal opinion, not a suggestion to buy. Please use this article as your DYOR resource for self-actualization or decision making.”

References:

https://docs.layer3foundation.org/

https://app.layer3.xyz/blog/reintroducing-layer3?slug=reintroducing-layer3

https://app.layer3.xyz/blog/introducing-the-infinity-cubes?slug=introducing-the-infinity-cubes

https://dune.com/l3/layer-3-overview

https://tokenterminal.com/terminal/projects/layer3

https://cryptorank.io/price/layer-3

https://defillama.com/narrative-tracker

NOTAI is a blockchain-based AI-driven super app project built as a bridge between Web 2 to Web 3. The platform is connected with AI to improve personal efficiency, financial management, education in web3 and decentralized finance (DeFi). NOTAI offers many features on its super app, AI portfolio management assistant, access to DeFi applications, fun Play-to-Earn games as learning facilities and openers in the crypto space. The two pillars that NOTAI holds are the use of AI technology for personal and financial growth, and financial independence through crypto education and DeFi exposure.

There are problems with the potential development of AI and Web 3 technologies that are challenging for ordinary people to access. NOTAI stands as a bridge and solution to these problems. In addition, some of the issues raised by NOTAI include DeFi complexity and usability, Security concerns in Web3, trust and familiarity, integration and interoperability in blockchain, economic bariers, bridging the digital divide.

Then, the problems that are often encountered as a medium to bridge Web 2 to Web 3 are the lengthy registration process and some require KYC, challenges with on-ramps, peer-to-peer transactions or entering funds with gas fees.

NOTAI is built on on-chain and Telegram which provides easy access, display, and gasless transactions. NOTAI, which is built on the Telegram platform also provides convenience for all Web 3 Finance utilities that were previously built into their own platforms (such as Aave, Compound, etc.).

NOTAI's mission is to bring 950M telegram users to crypto with AI. Telegram was chosen because of its ease in bridging Web 2 to Web 3. Telegram itself shows a very significant increase in the number of users every week.

In addition, TON managed to attract all crypto space interests to use the TON ecosystem both as users and developers in a saturated market phase. currently TON is still leading the narrative for Play-to-Earn airdrops with its Telegram mini-apps.

Token Utility & Tokenomics

NOTAI Super App offers untilities that support its vision and mission:

• Subcription model to access all AI products offered such as AI wallet analytics & abstraction, AI crypto score, AI news, etc.

• Gas-fees using the abstaction mechanism, making users pay lower fees due to simplified interactions on the blockchain.

• Community sales & launchpad allow users to access early-stage investment opportunities and locked tokens help reduce market volatility and encourage long-term and stable investment in the platform.

• Staking as an alternative income stream with the general mechanism that DeFi usually does.

• In Game currency and internal transactions, NOTAI tokens are also designed as in-game currency with various platform gamifications. Users can earn and sell tokens after successfully completing in-game missions or tasks.

NOTAI has already secured significant funding, raising $1.31 million from prominent investors like Ape Terminal, ChainGPT Labs, and Seedify Fund. The token distribution portion for private sale & public round is not as much as the token amount for public. NOTAI is committed to developing an ecosystem for the community.

NOTAI Super App

The AI portfolio assistant that NOTAI offers is generally the same as the DeFi + AI mechanism works, the difference here is that they are trying to bring the platform integrated with Telegram. By automating AI portfolio management, it makes the DeFi experience easier. At least what NOTAI reports on the official website is that AI in DeFi will include staking & farming, LSDFi & re-staking, bridging, lending/borrowing, launchapd, DEX, and other features.

AI Social is the perfect place to bridge Web 2 to Web 3 in the early phase. NOTAI incentivizes users to complete quests on a platform for gamification and education. AI Social will include Web 3 learning features, Quests, Clans, Referral Program, Rewards, Mini-Games.

AI app center is the subscription platform previously mentioned. All AI apps needed will be collected and connected in the AI App Center. Features covered are AI web3 assistant, AI wallet Analytics, AI portfolio manager, AI copytrading, AI games, AI web3 market insights.

In my personal view, NOTAI can be a good bridge to bring many new users into the crypto space. The future potential of TON and Telegram is also very good judging by the number of dapps deployed on the TON network and the increasing user activity. However, there are problems that NOTAI may face in the future, one of which is the super app mechanism with a subscription which is a common Web 2 business model and create a token Web 3, not all other projects do this. Of course, the price of subs has not been published because it is still in the development process, but if you look at the timeline for the future, it is clear that there are many AI data information platforms that are superior, sophisticated, and cheap for now. NOTAI's challenge is to create a unique selling point (UPS) for their subscription product.

NOTAI Now!

There are 3 phases that will be carried out by NOTAI, currently NOTAI is in phase 1. So it is true that for now NOTAI is not AI. Because now what can be done on the NOTAI platform on Telegram is mini-games and airdrop tasks.

On August 16, 2024 NOTAI launched the AI-curated web3 news platform. All curated by state-of-the-art AI, processing terabytes of data in real time to bring some information like breaking news, market analysis, and trend prediction.

Potential Competitor On TON Ecosystem

1. Ston.fi

Ston.fi is a platform built on the TON network that integrates AI for yield optimization. Ston.fi uses AI to move liquidity into various yield-farming pools to maximize yield. AI works to predict the best performing pools based on historical data, trends, and risk assessment.

2. DeDust

DeDust is a DeFi platform within the TON network that provides several AI-driven decentralized bots. DeDust's utility is to help manage assets, track portfolios, and trade on various platforms.

NOTAI Teams

Conclusion

NOTAI brings a good vision and mission for the crypto space, trying to attract many new users from Web 2 to Web3. The main products they provide also have a good impact on liquidity in crypto. Especially on the concept of bridging to Web 3 in the Web 3 learning product. I think it will be very useful for new users. Being built on Telegram also makes it easier for users to access and interface.

But there are some challenges that NOTAI may face, such as the main product which has a lot of competition for now. Also, competitors that are both built on Telegram have pretty good evaluation performance at the moment. NOTAI needs to re-evaluate the products they will build. how account abstraction works needs to be explained in detail to users in order to show the advantages of NOTAI over its competitors.

I personally still believe in NOTAI's product based on the information I have gotten. In addition, the very convincing team list makes me personally believe NOTAI can grow bigger in the future. NOTAI is just the beginning, there is still the next phase according to the roadmap and I look forward to good updates in the future.