Agent Naeem

Gm anon, are you paying attention yet? Principal engineer at Google tells us they worked on a distributed agent orchestrator for months. They then gave the problem to Claude Code, which reproduced their work and got them basically 80% there… in an hour. When building gets this cheap, what’s the moat? Ideas? Taste? Distribution? Speed? There is no better time to start experimenting & learning “vibe” coding. If you’re skeptical, do what @rakyll suggests and try it in a domain you actually understand, so you can judge the output. I made a multi agent orchestration app for blog posts last year and it blew my mind.

The deeper you get into a project (or vibe coding smth), the more context a fresh chat needs. That’s why I rate this @NotionHQ update. I’ve used Notion as my second brain for years, alongside Google Drive. Now ChatGPT, Codex and Claude Code can all work from and save to the same Notion pages. Say I use ChatGPT to research a feature and save the decisions we made and why... I can then get Codex to pick up the context and start building. When I inevitably hit my usage limit, instead of waiting for a @thsottiaux reset, I can switch to Claude Code and pick up ChatGPT's notes and where Codex left off. If the plan changes, the update goes in the same place for whichever agent I use next. If only you can see my context architecture in Notion btw. I been planning for this for a while. Organised chaos.... and it's about to get more chaotic. 😆

Crypto hands out the "AI agent" label like WL spots… and it's worth about as much in most cases. 😆 Telegram bot that buys when you type /buy? Agent. X account that posts Y every hour? Agent. NFT w/ 6551 + AI via MCP? Agent, obviously. By that standard I'm an agent too… it's literally in my username. 😂 None of this tells me what the AI model actually decides. "Swap 1 ETH for USDC" is a specified action… you don't need an agent for that. "Keep my stables earning decent yield without doing anything stupid" is a job but you could still hardcode that. Check 5 pools every Friday, move into whichever pays the best APY, with over $50m TVL… every rule is written in advance. The difference actually shows up when something happens that the rules didn't plan for. One pool's APY triples overnight… come Friday, the workflow apes straight in. An agent can stop, ask why, notice the incentives end next week and decide it's not worth it. That's discretion, that’s agency… and that's what makes it an agent. Tbh, for plenty of tasks I'd rather have the predictable automation. I don't need software getting creative with my bags… But I do want to know which one I'm trusting with my money: a process someone designed in advance, or a model making its own calls. TLDR: if every step is hardcoded and the model never chooses what happens next, you’ve not built an autonomous proactive agent 😤

Necessity used to be the mother of invention. Now, building's got so cheap that few stop to check whether anyone needs it. Building something used to cost enough that you only bothered when the problem/market fit was clear. The cost did the filtering, as nobody wanted to burn their time and savings on something nobody wanted. That filter's mostly gone rn due to AI… especially in the digital world. You can vibe code things (websites, apps, tools, tokens, NFTs) quickly… so the TL fills up with “solutions” wandering around looking for a problem, but often never find one. Tbf, cheap building is amazing as more people get to try things, and some of those hobby projects will turn into something real. But necessity used to do the filtering for free. Now it's on whoever's building… and judging by my feed, most people haven't realised. As building gets cheaper… knowing what's worth building becomes where the value is. 🤫

GM. Binance hack a bit sus ngl. More to it than meets the eye, I’m sure of it… But idk… mid terms incoming, war with Iran kinda expensive… what’s @SBF_FTX up to these days? 🤣

Anthropic's IPO filing is pretty wild tbh. The risk factors say concerns have been raised that sufficiently advanced AI could pose "catastrophic or existential risks to humanity" without the right safeguards. 💀 These AI models could help develop new models faster than they can be properly tested and kept safe. Flip back to page 6 and speeding up that research is part of the pitch. Each version of Claude helps build the next one. Anthropic calls it "closing the loop" and says it's "one of the most significant structural advantages in frontier AI". Over 90% of the contributions to its own codebase in September 2026 were written by Claude. If this starts moving too fast, who gets to say, hang on... we need more time? If you buy the shares, you're putting a lot of trust in the people already running it. The 7 co-founders keep voting control after the IPO, and a trust picks most of the board. I can see why one would keep those decisions away from short-term shareholder pressure... but the founders still have OpenAI and Google breathing down their necks. I'd want to know what happens when the safety team needs another month, and a rival's pulling ahead.

GM. @themutualfun is minting rn… pull up a chair, but don’t get too comfy. 🪑 I can see the appeal to mint at 0.01 ethereum:native ... but buying the token or NFT to fund someone else’s exit? Less convinced. 1. THE IDEA: 5 Funds. 800 NFT Seats in each. Pay $TMF to enrol, then vote weekly on investments and payouts. The menu? 10 Robinhood stock tokens and 9 crypto tokens (incl. $STONKBROKER & robinhood:0xc72f232a6869e6cf34dc06129affd07f8a2a246a )… unallocated money sits in Ethereum. Up to 30 ETH from the mint go to funds. Members can receive voted share of profits, paid in the assets the fund holds... separately, 30% of each fund's $TMF fee stream goes to enrolled holders in TMF. So some profits come from trading activity, just like StonkBrokers. @Kingpickle also did some last minute expectation setting, telling people not to expect huge payouts. The aim is to build the “collective pot”. 😆 2. THE CHAIRS ARE ALREADY BAGS: 4,001 total Seats. 3000 available through Passes, 1000 in Anvil liquidity and 1 Chairman's Seat. Each Pass buys 2 Seats at 0.01 ETH each. Mint allocation distribution: - up to 1,648 through holders of the founder's previous collection, Dengs. - up to 1352 through UGC applications, collabs, team reserves and invites. Who Seats have gone to will be key in the success of this project. A Seat can be swapped for 368,920 TMF tokens once trading opens. Founder also said buy the token and “get a fat bag”. But minters don’t need to... the chair is the bag. 3. FOLLOW THE EXIT: 75% of TMF tokens are available to cover Seats being sold to Anvil. 25% is allocated to public launch. 3,000 minted Seats = 1.1B TMF, 2.76x the entire launch allocation. Some will keep their Seats… others will need TMF to buy Seats and enrol (with half enrollment fees burned. But every pump of the token makes it more tempting to cash out the 0.01 ETH Seat. The reserve can honour the token amount while the market murders its ETH value... Will enough people want to join and stay to absorb those who want to leave? Rhetorical… and a fixed conversation rate doesn’t answer this. Also... for up to 180 days, the Fire warden key can pause the market immediately… with an hour public notice it can move the reserve. Quite a lot riding on the fire warden behaving himself... 4. AFTER THE LAUNCH PARTY: Funds investment profit payouts starts at 0... and at least 25 seats must vote to change weekly settings. Low voter turnout leaves the previous settings running. Then comes week 52. Each fund closes unless "continue" beats "close" with at least 25% eligible voting weight quorum. A tie closes it. A small committed group could keep a fund alive… and closure pays the remaining holdings to enrolled members by weight. My call? Ghost office within 6 months... and I'd be very surprised if all 5 funds survive their first Annual Meeting. A 0.01 ETH mint + enrollment for an investment DAO could be a reasonable short term punt. Anything more than that, including buying a "fat token bag" while thousands of cheap chairs carry their own exist is a different proposition. Can it pump? wouldn’t surprise me during the honeymoon period. After that? down and to the right, with office clearance prices. 📉 See you at the Annual Meeting.