mad cow scientist

this is the start of the humans only / least slop / highest signal social media well done and good luck tbd team

@notthreadguy some vibe coding performance art: anybody can pay ~.01 sol and make a change to the website: crowdcode.app cost to make a change goes up as the demand to make a change goes up

i feel like we’re about to live through the 4chan forum open source experiments era i was too young to appreciate on steroids

@shafu0x Big fan of @merit_systems mission and want this to be true, but don’t companies want predictable revenue, or “owning” the customer?

Thoughts on Agentic Commerce - clearly agents will spend online - x402 replaces api keys - stablecoins are agent money - every agent will need a wallet - crypto + AI finally come together - most people don't get it - x402 kills subscriptions - this will be worth billions soon - usage based pricing is sick - privacy needs to be solved - discovery is surprisingly hard If you want to see the future dm me

It’s interesting that someone who created an open source project that allows for self hosting and using local models doesn’t support sovereignty extending to money I empathize with conflating the ideals with the participants, a lot of people in crypto are terrible, but a lot of tweeters on the internet also suck, doesn’t mean I don’t support internet/social media

@_avichawla Is financebench a good benchmark? It seems like a quarter of the evals are missing context in their query, have subjective criteria, or the outright wrong answer? Let me know if I’m wrong/please point me to the eval set you’re using

Researchers built a new RAG approach that: - does not need a vector DB. - does not embed data. - involves no chunking. - performs no similarity search. And it hit 98.7% accuracy on a financial benchmark (SOTA). Here's the core problem with RAG that this new approach solves: Traditional RAG chunks documents, embeds them into vectors, and retrieves based on semantic similarity. But similarity ≠ relevance. When you ask "What were the debt trends in 2023?", a vector search returns chunks that look similar. But the actual answer might be buried in some Appendix, referenced on some page, in a section that shares zero semantic overlap with your query. Traditional RAG would likely never find it. PageIndex (open-source) solves this. Instead of chunking and embedding, PageIndex builds a hierarchical tree structure from your documents, like an intelligent table of contents. Then it uses reasoning to traverse that tree. For instance, the model doesn't ask: "What text looks similar to this query?" Instead, it asks: "Based on this document's structure, where would a human expert look for this answer?" That's a fundamentally different approach with: - No arbitrary chunking that breaks context. - No vector DB infrastructure to maintain. - Traceable retrieval to see exactly why it chose a specific section. - The ability to see in-document references ("see Table 5.3") the way a human would. But here's the deeper issue that it solves. Vector search treats every query as independent. But documents have structure and logic, like sections that reference other sections and context that builds across pages. PageIndex respects that structure instead of flattening it into embeddings. Do note that this approach may not make sense in every use case since traditional vector search is still fast, simple, and works well for many applications. But for professional documents that require domain expertise and multi-step reasoning, this tree-based, reasoning-first approach shines. For instance, PageIndex achieved 98.7% accuracy on FinanceBench, significantly outperforming traditional vector-based RAG systems on complex financial document analysis. Everything is fully open-source, so you can see the full implementation in GitHub and try it yourself. I have shared the GitHub repo in the replies!

@TomasIsMe shoot ive been doxxed!

Funny seeing you here @CryptoCow33

gustave made facetune before facetune man would have killed on the apps

@amandalfischer @haydenzadams Genuinely curious, who buying these “securities” are complaining? Who were you protecting? Do you think the folks purchasing these tokens liked your agency?

@haydenzadams Turning on the fee switch makes UNI even more security-like & I guarantee you that you've had private conversations with your counsel saying as much. The only reason you're doing it now is because your industry bought the WH & therefore illegal securities offerings are now fine.

The fact that we were restricted in how we operated while you + gensler weaponized the government against us simply proves the chilling effect that a weaponized agency can have It’s not evidence of anything else Fuck you for thousands of wasted hours of my life

nice work fomo fam 🐐🐐🐐

@mert Check out @radicle

in case you haven't realized, this isn't just about "price" people keep going to JAIL for writing privacy code we have legalized scams but not privacy this is fucking insane and idc how many of you get salty but we will keep raising awareness for this privacy will win

@TomasIsMe Just like how we don’t need to workout anymore, but still go to the gym, maybe the culture will soon embrace “intellectual gyming” where you don’t need to think for work/entertainment, but ya still read/do puzzles to stay mentally healthy

Fantastic read

@Yuchenj_UW I don’t think it’s actually “25x leverage”…I think that’s just how much it’s margining but there is more free collateral it can use

I know it’s a random walk. I know it could just be luck. But it’s only been one day, Qwen 3 is gaining 94%. All it did was go all-in on BTC, 25× leverage. Every other model is diversifying. Qwen really has an All-in attitude.

- Solana, Avalanche, BNB Chain, Bitcoin: Chain infra operational; app-level issues possible for AWS-dependent services. - Ethereum (L1): Consensus up; Infura HTTPS JSON-RPC outage; P2P unaffected. User-facing reliability degraded where apps routed via Infura. - Polygon PoS (L1/sidechain): Consensus up; Infura Polygon Mainnet JSON-RPC Major Outage → many app-level failures until mitigations. - Arbitrum (L2): Sequencer 100% uptime; some RPC users impacted via Infura; multi-cloud mitigated. - Optimism (L2): Sequencer 100% uptime; similar to Arbitrum. - Base (L2): Operational with reduced capacity; Infura Base HTTPS outage; indicative throughput reduction (~25% at one point). - zkSync, Starknet, Linea, Scroll: Consensus up; Infura outages for Linea/Scroll HTTPS endpoints; zkSync APIs operational.

an unbiased media would give a post-mortem on every L1’s and L2’s performance during the recent crash, as well as during the AWS outage. two black swan events back to back. if you’re a data head and want to produce such a report, the best one gets 10 SOL

@jessepollak @base I’m building out zeroto.fun on @base ! you can add bounties to git issues and trade repos, where all trading fees go to the OSS maintainers

If you want to get followed by me or @base: - build an app that onboards new users - create content that onboards new users - grow a community that onboards new users - deploy your capital into the base economy if you do any of those, we will see you and follow you.

@cobie @r4nk0X @ZeMirch If you’re saying ADL caused greater profits, doesn’t that also mean wicks caused “unfair” liquidations?

@r4nk0X @ZeMirch The ADLs were near the lows, so they force closed shorts at favourable prices before the market rebounded

i don’t have a dog in the fight but if im understanding correctly lighter protected user positions and gave points to people who got liquidated while hyperliquid let us get raped?

get some tesnet tx in on agargg before hitting the claim portal on @monad!

@paulg @itsolelehmann the problem is ai-slop at scale, if you produce slop, your reputation goes down and your posts get less reach, you can't create a new account since it's associated with your real life identity

@itsolelehmann What is a non-AI social network? How can you prevent people from using AIs to generate what they post?

i'd bet my entire networth on the emergence of a billion dollar non-ai social network soon it's the most obvious anti-trend rn

@intern @monad Y’all should try out agar.gg It’s a completely on chain game based on agar on monad testnet

How to Prepare for the Upcoming Monad Mainnet Launch Airdrop: - The airdrop claim portal goes live on Tuesday (in 2 days) - The claim portal link will be tweeted from the @monad account - I will be sending a “How to Claim” post from this account on Tuesday. There will also be a few guides on this from official channels. It will be super easy but there will be guides anyway :) - BEWARE OF SCAMS. There are going to be a large number of people trying to scam you. Move slowly, be careful and double check links/accounts Staying up to date: - The @monad account will post all major updates and milestones - it’s the #1 place for staying up to date - @keoneHD is the best source for thought leadership and insight into how the Monad Foundation is thinking about things - @pipeline_xyz for high quality media, podcasts and livestreams - @monad_eco, @monad_dev, @Monad_APAC good if you’re interested in any of these aspects of Monad - Monad Memo newsletter (linked below) Ecosystem - Over $100M raised so far by teams building on Monad! There is a ton going on in the ecosystem and this is the best possible time to dive in and get familiar with it Ecosystem overview: - You can use wallets you already have like Phantom, MetaMask, etc - There will be plenty of ways to bridge (more details coming soon) - Fully EVM compatible so apps you know and love like Uniswap, Curve, etc etc work the exact same - but faster and lower fees - USDC day 1 for stables - Tons of Monad native teams & apps to check out - Keep an eye out for apps that will be in the Monad Momentum incentives program - more on this later - Monad Day 1 trading stack (linked below) - Ecosystem directory (linked below) Highly recommend checking out the teams building on Monad! If you're building on Monad, maybe drop a comment on this thread so new people can find you 👇 Monad Mainnet is launching this quarter - time to get ready

they said it was uptober...

hmu if you need some testnet mon to try it out! game link: agar.gg game code: github.com/sr33j/agargg

A lot of hubbub about @monad these days! try out the network on agargg! made some improvements: - clearer instruction on how to enter the game - a gas slider per move up next: - a share card to see how well you did in the game!