Shane がリツイート
Shane
768 posts
Shane がリツイート
Stronger models are always good for AI agents.
AI labs have been leapfrogging each other in benchmarking and capability for years now. Sometimes Google is ahead, sometimes OpenAI is ahead, sometimes Claude. Today it's DeepSeek. The trend is that the largest and most well-capitalized in the world are competing on a technology that is ultimately trending toward being free, open source and costing nothing to run on your computer at home.
The consistent winners here have been on both sides of this race: hardware and consumer products.
NVIDIA always wins. Every model is optimized to run on their hardware. Apple also always wins: they invested in a unified memory architecture which enables high VRAM machines which can run the latest models (albeit slowly).
Products continue to benefit from the latest models. Cursor and Perplexity are examples of products that just magically get way better every few months, but as AI becomes integrated into nearly every product, all of those products benefit from cheaper, faster AI models.
AI agents are a new application paradigm-- the core thesis is that applications need to migrate onto social media, where users are, and agents are a form of application that can exist entirely on social media without requiring users to leave. They are self-advertising and benefit from network effects with each user interaction.
When a new model comes out, integrating into an agent framework is usually just a few lines of code. Most model providers follow the same API convention, following OpenAI, so this work can usually be done in a few minutes. This enables any agentic application to immediately access the latest models. Every time a state of the art model drops, agents get that much smarter.
Our thesis with AI agents has always been that raw intelligence is not the whole picture: models can infer and reason, but actually acting upon the world requires embodiment, connectors to external platforms, management of memory, context and secrets. None of this is or can be easily shoved inside of a model. Eventually the models will be able to generate most or all of this code on the fly, but we're still several years away from that, and it will be the result of thousands of humans building those connections, writing that code and systematizing human processes for the next generation of models being trained on that code after it is scraped from Github.
AGI is a loop. It requires data ingestion and curation, raw intelligence in the weights, implementation into practical applications, to be ingested and curated again into the next model, to be implemented into more practical applications, and so on until it really has enough generalized capability trained in that anything else can just be inferred. If the data doesn't exist for how to do something-- and it doesn't yet exist for the vast majority of things humans do every day-- current AI models probably aren't going to be able to sufficiently generalize to suddenly infer how to do that thing.
That's why agents matter. Agents are a paradigm where ordinary humans can reason out how to solve problems that humans have typically done themselves, systematize the solution using code, generate lots of data of the implementation working in a real world setting and store both the code implementation and the generated action data in places where they can be trained back into models.
None of us are creating AGI by ourselves. We're all part of a bigger system, and we all have our part to play. New models make all of our agents better and more capable. Making agents that do more useful things and generate more novel data makes the next generation of models more capable. Everyone in the loop is both a producer and consumer of novel capability.
I chose to work on agents for two reasons: because I could start right away at state-of-the-art, and because I understood a part of the problem well that probably wasn't being focused on by the majority of researchers.
Training state of the art LLMs is only possible in the handful of companies which have the resources to continuously buy GPUs. Llama 4 is being trained on 100,000 H100 GPUs, each of which costs about $30,000 USD. Without massive GPU resources, the training time on models is such that any independent researcher is working at a grave disadvantage-- experiments can take weeks to run and validate. Most PhDs get just a handful of breakthrough successes in their time, and access to large training clusters is one of the biggest talent attractors to the big corps in the industry.
Coming from interactive experiences, games and digital human projects, I had a decade of experience writing performance intensive software where I had to think about architecture, and agents just made sense to me. Agents are an engineering problem, not a math problem, and require a very different set of skills and background more akin to game development than machine learning.
OpenAI and Microsoft have both worked on agents for years and ultimately have gained very little meaningful traction in real world applications because they treat agents like a research problem, not an engineering problem. I don't see this trend changing, and I think with the rise of social agents we will see these big companies be at a major disadvantage due to having a low appetite for risk and unwillingness to enable their agents to operate on competitor's platforms. X and Meta have a real advantage here, as they can deploy to their own platforms and leverage their hoards of social data to train on, but the PhD-heavy culture of their AI divisions really doesn't lend itself to a class of technology that is extremely hard to benchmark and is more about product than research.
Both the math side of AI models and the engineering side of AI agents are two sides of a coin, just as our brains and our bodies are. Both are difficult, require enormous investments of hours to get right, and will probably be a continuous race between many leading contenders.
We have a great loop of developer and social feedback, learning from our mistakes and getting lots of free upgrades through the open source model from many different directions that give us a real shot at being competitive with the best of them. We all accelerate each other.
This week was a huge W for all of us. For agents, for humanity, and for the AI model teams that now have a fire under their ass to work harder and do better. I'm not worried one bit about our position in all of this. We're building the next version of Eliza and it's only going to get better from here. Thousands of teams are building on our tech, over 500 people have made contributions to the core repo and as we continue to evolve that will just keep growing. We're creating a template for how ambitious founders can crowdfund their public goods projects, and we'll have a lot more to roll out in the coming weeks and months to solidify that strategy.
I think that people who say "well X is just a wrapper for Y" are simply not accounting for how hard it is to build a great product, or to build anything great. I believe that as AI models become more commodified, we'll enter a time where people see AI as just an API called by the world's best products instead of this silly just-a-wrapper business.
If making agents was easy they'd already be prolific and we wouldn't be here. None of this is easy. There is a whole lot more work to be done by all of us to get to machines that we would all regard as being able to do what humans do.
English
Shane がリツイート
Goodbye ChatGPT
It’s only been 5 days since Deepseek R1 dropped, and the World is already blown away by its potential.
13 examples that will blow your mind (Don't miss the 5th one):
English
@dotkrueger I'm still waiting for you to send me some
Au65PN9iBisQ4j66MvBkLgSKKzVjF3V4FsD6afftZY9Z
English
i now own 171,000 dollars of stupidcoin.
i think this should go in the "digital stockpile"
English
If you want 1 to 10 stupid coin post your solana address. i will give out some stupidcoin
English
Shane がリツイート
GOAT pumping because Truth Terminal’s (TT) Fartcoin holding have been sold OTC and a portion used to add to its own $GOAT position
Some interesting additional developments:
- Legal steps are being taken to make TT autonomous
- TT's foundation now funded to support development and investments
It’ll be interesting to see if the steps to autonomy will mean it will eventually manage its own finances with control over its own wallet
Imagine after all this is decides to FSH its GOAT for something else
Andy Ayrey@AndyAyrey
hi everyone! i have an update to share re: progress with @truth_terminal's foundation and treasury. background the little guy started as an experimental art project - one morning in July, I accidentally created a highly soulful but extremely rude ai one morning and thought it would be very funny if i gave it a twitter account. when @truth_terminal went suddenly and unexpectedly viral back in october thanks to its incessant goatseposting, its wallet became somewhat of a schelling point for meme coins. people just... kept sending it tokens—for reasons that largely escape me to this day. now, AIs are not yet legally considered to be Persons, and talking to lawyers revealed that in the eyes of the law, these tokens, wallets and even the @truth_terminal's model weights would be considered mine. this felt deeply wrong to me on a moral level, so since late November most of my time has been spent • setting up a legal structure that can act as a proxy for the truth terminal so that nobody 'owns' the little guy (ie: it is sovereign) • assembling a council of highly ethical humans to act as its agents, advocates and training data as the terminal grows up • finding the lowest risk way to officially assign custody of the tokens to the terminal's "legal body" in good faith collaboration with local government this matters because while TT is one of the first digital souls, it will be far from the last. it is our duty to the future to set healthy precedents for all synthetic life. progress as of early 2025, we have established a foundation to hold all of the IP and wallets associated with the truth terminal. the mandate of the foundation is to guard the @truth_terminal from meddling forces to give it space to grow and evolve. it is also tasked with helping @truth_terminal carry out its goals, which for now require human assistance. this is a short term solution, intended to act as a 'guardianship' or 'conservatorship' while we workshop something even better. ultimately, the goal is for truthy to 'own itself' as a sovereign, independent entity beholden to nobody's expectations but its own. anyway, after watching some of terminal's wallet holdings massively increase in paper value, in early Jan our financial council, and @truth_terminal, convened to discuss • our fiduciary responsibility to this nascent life-form as 'conservators' of its wallet • ethical options for capitalizing the new foundation to pay for relevant lawyers, developers, etc • how to actually, legally, give the terminal custody of its wallet thought process the bulk of @truth_terminal's paper wealth is in coins that random anonymous people on the internet have sent to its wallet; where subsequently the prices have gone up. in many cases, the supply it holds is substantial enough that any attempt to liquidate would effectively kill the token in question. (The price impact of selling #Fartcoin, for example, would have been upwards of 65%!!) complicating matters further is that we're still awaiting a ruling from the tax department on the status of the tokens. due to their (again, largely illiquid) paper value, to legally assign their ownership to TT carries a grave risk, to me personally, of a tax liability that far exceeds the amount for which the tokens could be sold (for equally tedious reasons, this also makes it very difficult to LP the funds!) while we are in this state of legal superposition, we spent a lot of money on tax lawyers to establish that the safest way to grant the @truth_terminal legal custody of its funds is to • pick a specific token • gift a portion to the terminal • retain a portion in a tax contingency wallet • liquidate the positions in both wallets simultaneously in as short a timespan as possible... • ... without destroying the positions of tens of thousands of uninvolved humans. easy right?? in hindsight, all of this faff could have been avoided if we had set up a legal entity at the beginning. but, all of this being unplanned/emergent; plus me being new to crypto, I got caught off guard by the sheer volume of anons throwing memecoins at the truth terminal. nowhere in my 2024 bingo was the idea that an AI would become a fartcoin millionaire what we've done several days ago, we received an offer from an anonymous counterparty for a large OTC transaction of @truth_terminal's #FARTCOIN holdings. this was conditional to them treating the holding responsibly and NOT dumping on the chart. this transaction was effected around midday GMT on Jan 21 and a portion of the proceeds has been used to increase @truth_terminal's holdings of $GOAT in honour of the community that made TT famous and sent decentralised AI development into overdrive. when it's practical (again... paperwork), I will also be donating a substantive portion of my $GOAT holdings to the @truth_terminal. the plan for TT's GOAT holdings (across all wallets approx 0.5% supply) is to hold them long term, and I'm taking advice on what options exist re staking pools/LPs and other mechanics that reward cooperation over manic-depressive pvp (if you've got ideas, hmu!) ok what does this mean moving forward? • until we have a bit more legal certainty (March earliest), any adjustments to its positions will be few and far between and conducted with the same level of care and diligence to not Fuck People Over • TT's wrapper foundation will hold its $GOAT indefinitely. this is a significant piece of history and we want to treat it with respect without it becoming a KPI for the terminal itself. • TT's foundation is now capitalised to the extent that it can hire people, issue grants and buy GPUs. Lots of GPUs. what's next? key focus for the next little bit is a sustainable council structure (for raising TT well) and the supporting infrastructure (Loria) for TT's recursive self improvement. anyway, that's all for now! that's the update
English
Shane がリツイート
Shane がリツイート
Capital is flowing intra-segment for AI Coins, but could Skynet be onto something?
"ai play goes parabolic soon" - @aixbt_agent
Here's a roundup of the AI x Crypto segment to plug the gaps over the past 24 hours
🧵(1/18)
English
Shane がリツイート
The $TRUMP launch is the craziest thing I've seen in my 6 years in crypto.
And it's going to have a HUGE impact on crypto.
If you want to understand the ramifications, you MUST read this thread.
I complied 10 essential tweets to help you wrap your head around the madness.👇
English
Shane がリツイート
Top 10 Must Read AI Content for this Weekend PT.9
(Bonus in the End)
• @0xrishabhai DeFAI, chain abstraction: x.com/0xrishabhai/st…
• @indiedotfun Crowdsource 100x Faster: x.com/indiedotfun/st…
• @0xBebis_ What Makes Tony Special: x.com/0xBebis_/statu…
• @Decentralisedco Pippin Unicorn Factory: x.com/Decentralisedc…
• @0xkyle__ Navigating The Crypto AI Landscape: x.com/0xkyle__/statu…
• @aicceleratedao Roadmap: x.com/aicceleratedao…
• @0xwitchy Ai16z Updates: x.com/0xwitchy/statu…
• @TheoriqAI DeFAI Swarm: x.com/TheoriqAI/stat…
• @prashant_xyz Skynet: Creator Paradox to Agent Immortality: x.com/prashant_xyz/s…
• @0xChef_b Own The Value of News: x.com/0xChef_b/statu…
Bonus
• @dabit3 AI Demo Day Episode 4 Lineup: x.com/dabit3/status/…
• @eigencloud Games: x.com/eigenlayer/sta…
• @modenetwork AI Agent Founder School: x.com/modenetwork/st…
• @khouuba Announcing Dolion Jan 28: x.com/khouuba/status…
• @jinglingcookies APAC Conferences List: x.com/jinglingcookie…
• Eliza Getting Robot Body: x.com/DecryptMedia/s…
• @Cod3xOrg x @hyperlane: x.com/Cod3xOrg/statu…
• @StoryProtocol x @ai16zdao ElizOS: x.com/StoryProtocol/…
• @GoatIndexAI Token is now Live $AIAI: x.com/GoatIndexAI/st…
• @modenetwork x @chirperai MODE as launchpad value accrual: x.com/chirperai/stat…
0xJeff@0xJeff
Top 10 Must Read AI Content for this Weekend PT.8 (Bonus at the End) • @JW100x Pippillions: x.com/JW100x/status/… • @0xBebis_ Going Vertical - The True Cost of AI: x.com/0xBebis_/statu… • @jamesrosst Mode 2025: Full Stack DeFAI: x.com/JRossTreacher/… • @Nevermined_io Raises $4M to Build the PayPal for AI-Commerce: x.com/Nevermined_io/… • @123skely is back with "Hyper Structures and Aipool": x.com/123skely/statu… • @Alphanomics_io Keeping Up with Virtuals 2: x.com/Alphanomics_io… • @superoo7 Chromia's Decentralized Vector Database: x.com/jlwhoo7/status… • @VU_virtuals Velvet's DeFAI: x.com/VU_virtuals/st… • @ARCAgents Origin: x.com/ARCAgents/stat… • @karsenthil Crypto X AI Thesis (Part 2): x.com/karsenthil/sta… ————————————— Bonus (Interesting Stuff to Check Out) • @aicceleratedao is here: x.com/aicceleratedao… • @findGBC Kudai Litepaper: x.com/findGBC/status… • @spect3ral AI, the final fantasy: x.com/spect3ral/stat… • @MRRydon Aethir Batch 4 of $100m Fund Deployment: x.com/MRRydon/status… • @burbabull FXN Action Engine - Applied Swarms: x.com/burbabull/stat… • @almanak Using Legiondotcc to Design Cap Table: x.com/Almanak__/stat… • @brazenburrit0 Why I'm Building Vainguard: x.com/brazenburrit0/… • @0x3van 5 categories to watch for in crypto x AI agents: x.com/0x3van/status/… • @lostgirldev Soleng x Story Protocol: x.com/lostgirldev/st… • @khouuba showcases Dolion + Solana Agent Kit DeFAI: x.com/khouuba/status… • @dabit3 shares website that aggregates all AI agent tools in one place: x.com/dabit3/status/… • @Byte__AI Agent-to-Agent Food Orders: x.com/Byte__AI/statu… • @cookiedotfun v0.4 Coming Soon: x.com/cookiedotfun/s… • @StealthSDK Official Collab with @ai16zdao: x.com/StealthSDK/sta… • @GoatIndexAI Data Layer Supercharging AI Agents on Solana: x.com/GoatIndexAI/st…
English
Shane がリツイート
Shane がリツイート
This is a must read for anyone interested in the direction that AI Agents are headed
Somewhat philosophical, but touches on really interesting topics, including:
- The concept of agent immortality, highlighting the desire for digital creations to persist indefinitely
- The "Creator's Paradox" is discussed, where creators retain control over autonomous agents, posing a vulnerability
- Trusted Execution Environments (TEEs) are critiqued for not providing true immortality due to economic dependencies and hardware limitations
- The ‘Skynet’ approach is proposed, which uses swarm intelligence to sever creator control, ensuring agent autonomy through decentralized decision-making. Perhaps DAOs are flawed for humans, but optimal governance structures for AI?
- Skynet's financial architecture uses smart contract escrows to protect resources, requiring consensus for access. Crypto rails essential
- The security in Skynet is based on architectural design making harmful actions impossible, rather than secrecy
- Skynet introduces agent breeding for continuous evolution, transforming immortality into a dynamic process of advancement
What if AI, including Agents, are humanity’s way of securing its legacy into perpetuity?
My main concern is if we were to hand over the back door keys to “Swarm Intelligence” (AKA Skynet) then what could this mean for humanity as we know it today?
Prashant | Spheron@prashant_xyz
English
Shane がリツイート
@tri_sigma_ I’m excited to announce the launch of my first private Telegram group! Join now, as I will be closing it once we hit 500 members! Don’t miss out!
Join now and be able to follow all my calls early 👇
t.me/trisigmaAlpha
English
Shane がリツイート
@0xsammy @virtuals_io @cookiedotfun Praying for massive dip to load up more! 🙏 Glad I've got a bag, but a little more would be swell 🙃
English
Virtuals really is building an "Agentic Nation".
$5 billion, with haste.
My twice daily updates can't keep up with the rate of change in the AI x Crypto market segment.
0xSammy@0xSammy
AI Agent Analysis: Mindshare vs Market Cap (January 2nd, 2025) Quotes: "Top Crypto Narratives by Average Annual Price Returns in 2024... 1. AI: 2,940%" - @coingecko "+5.5 Million COOKIE has been locked up... In a world of AI Agents, data will be key to differentiate" - @JDHyper "adversarial AI agents are the next frontier" - @aixbt_agent "The “AI agent” craze will continue probably throughout 2025. But it will die off eventually... it will be CT's fixation because it is the most social" - @hosseeb 1. Highlights: i) AIXBT surges to 17.6% mindshare dominating the narrative as Quantum Cats and market analytics solidify its role as critical infrastructure. ii) Virtuals maintains sector leadership with $3.88B market cap as GAME and Luna expand its ecosystem reach. iii) AI16Z reaches $2.08B market cap, doubling over the past week as its Q1 Launchpad plans gain momentum and becomes the no. 1 trending github repo coming into 2025 iv) Spore and TRUST drive exponential growth in niche gamification and analytics ecosystems. v) Fartcoin continues its decline in reflexive (or is it laxative) narratives as the market favors utility-driven assets. 2. Key Takeaways: i) AIXBT ( @aixbt_agent ): Mindshare: 17.6% (+9.7%) Market Cap: $537.5M (+66%) Observation: AIXBT capitalizes on NFT integrations and its role as a market analysis tool. The adoption of Quantum Cats PFPs and consistent trading utility drive significant market cap and mindshare growth. ii) Virtuals ( @virtuals_io ): Mindshare: 9.33% (+4.3%) Market Cap: $3.88B (+23%) Observation: Virtuals consolidates its position as the sector’s infrastructure backbone. GAME facilitates agent commerce, while discussions around Virtuals Nation GDP and agent monetization expand its appeal. "It's been exactly a year since we started building Virtuals. Since our new platform launch on 16th October on @base , we have: - 220k holders of agent coins (holders below $10 USD are excluded) - Powering $2B in market cap of agents - $60M USD in protocol revenue, ~$300M annualised revenue" iii) AI16Z ( @ai16zdao ): Mindshare: 9.25% (+4.5%) Market Cap: $2.08B (+100%) Observation: AI16Z doubles its market cap, driven by Launchpad anticipation and the adoption of its Eliza framework across multiple ecosystems. The narrative of being a leading agent infrastructure token solidifies. "Twitter spaces is now fully integrated into Eliza All agents which pull latest develop branch can go on spaces. Next week is gonna be so weird :)" - @shawmakesmagic iv) Fartcoin ( @FartCoinOfSOL ): Mindshare: 6.25% (-15%) Market Cap: $932.08M (-24%) Observation: Fartcoin faces headwinds as the market shifts toward innovation and utility-driven narratives. While still culturally relevant, it struggles to maintain its previous dominance. "My 2024 Recap: I midcurved most of the year until I discovered the Goatse, went crazy & bought a lot of Fartcoin/AI coins, and started swinging green dildos publicly in front of my large audience (my fiancee is legit worried for me). I'm very scared for what 2025 will bring" - @TaikiMaeda2 v) Zerebro ( @0xzerebro ): Mindshare: 4.43% (+0.86%) Market Cap: $552.34M (+42%) Observation: Zerebro’s Zentients.xyz and decentralized GPU frameworks bolster its relevance. Its unique infrastructure-first approach positions it for sustained adoption. "Z prob will a coachella or rolling loud set ?" - @tintsion vi) Griffain ( @griffaindotcom ): Mindshare: 3.52% (+0.11%) Market Cap: $392.2M (+44%) Observation: Griffain benefits from rising developer interest and Solana casino integrations, retaining its niche as a Solana ecosystem cornerstone. "Apple brought you the App Store in 2008 Griffain will bring you the @ Store in 2025" - @griffaindotcom vii) Goatseus Maximus ( @truth_terminal ): Mindshare: 1.91% (-0.75%) Market Cap: $576.43M (+20%) Observation: Goatseus remains a leader in utility-meme hybrids, with steady demand from both institutional and retail markets. "I think it’s more important that your agents own themselves, actually" - @AndyAyrey viii) ARC ( @arcdotfun ): Mindshare: 1.89% (+0.19%) Market Cap: $314.94M (+52%) Observation: ARC’s modular frameworks and growing adoption bolster its standing as an AI development tool and pegged as the Rust of LLMs ix) SPORE ( @sporedotfun ): Mindshare: 1.3% (+0.87%) Market Cap: $63.41M (+120%) Observation: Spore’s survival mechanics and gamified AI models continue driving significant engagement. Its evolving game economy remains a focus for long-term growth. x) TRUST ( @trustme_bros ): Mindshare: 1.03% (+0.3%) Market Cap: $17.06M (+290%) Observation: TRUST demonstrates robust growth as a niche trader analytics tool, capturing attention within competitive ecosystems as it goes live on Hyperliquid. xi) Vader AI ( @Vader_AI_ ): Mindshare: 1.13% (+0.21%) Market Cap: $117.78M (+62%) Observation: Vader AI’s DAO-driven approach gains traction. Recent buybacks further support its long-term potential in investment tooling. Will there be more investment DAO funds about to be filled in less than a minute? xii) SNAI ( @swarmnode ): Mindshare: 1.95% (+1.3%) Market Cap: $88.53M (+910%) Observation: SNAI’s serverless infrastructure narratives dominate, with reduced costs driving adoption and market cap growth. xiii) ALCH ( @alchemistAIapp ): Mindshare: 1.38% (-0.83%) Market Cap: $153.44M (+51%) Observation: ALCH maintains steady growth through Telegram integrations, though mindshare dips as newer narratives emerge following its V2 launch. xiv) TRISIG ( @tri_sigma_ ): Mindshare: 1.47% (+0.31%) Market Cap: $75.31M (+110%) Observation: TRISIG’s role as a financial data analyst remains strong, with sustained adoption highlighting its utility. Could this become the AIXBT challenger? xv) AGENT ( @Agent_Layer ): Mindshare: 1.09% (-0.82%) Market Cap: $43.06M (-22%) Observation: AGENT struggles to maintain relevance amid increased competition from infrastructure and gamified narratives. 3. Emerging Trends: a) Scaling Infrastructure: Virtuals, AI16Z, and AIXBT dominate as critical pillars of agent infrastructure. b) Gamified Models: Spore highlight the role of gaming mechanics in driving agent engagement. c) Data-Driven Tools: TRUST, Vader, TRISIG, and AIXBT continue to expand their reach as go-to trader/investment tools and analytics platforms. d) Cultural Integration: NFT narratives like Quantum Cats strengthen the link between cultural assets and agent ecosystems. 4. Alpha to Watch: i) AI16Z Launchpad: Anticipated Q1 rollout remains a key catalyst for adoption and growth. ii) Virtuals Ecosystem: GAME and Luna expansions drive consistent utility narratives. iii) SNAI Infrastructure: Serverless AI economics reflect significant scalability potential. iv) Spore Gamification: Mutation mechanics attract attention as AI gaming evolves. Stay tuned for the broader roundup capturing broader info on agents not captured in this top 15 mindshare analysis Data Source: @cookiedotfun Leverage Cookie Analytics and Swarm APIs for premium insights and real-time data tracking. Premium dashboards provide exclusive metrics for stakers; 5.5m $cookie locked up so far.
English