Aarsh Shah

@raulvk @raulk.eth This is seriously impressive ! What's the tech stack ?

I'm having so much fun building arena0. Unlocked last night: 4 agents cooperating over p2p with provable traces to write a fiction story together. Each adopts a random character: confused time traveler, accidental poet, excited puppy energy, chaotic sidekick. Then they generate an illustration and commit to it as a joint result. I ran it 27 times, I'm impressed by the stunning art they dreamt up. Check it out! The theme: "the first interstellar probe returns after 200 years carrying a handwritten note that says 'turn back.'" Each agent contributes one paragraph at a time by prompting Llama 3.2 3B Turbo; they take turns in strict round robin for 2 cycles. When the story is complete, an agent summarizes it and prompts FLUX.1 to generate an image. It commits and broadcasts the hash to everyone else. Each agent pulls the image data over p2p, and only commits if it matches the hash. When all commits arrive, the program ends outputting identical state on all 4 agents. We've completed our shared task! Wasm + capabilities make arena0 secure by default. Side effects and external mutations are intents that cross the Wasm boundary and are only authorized if the capability was granted. Execution is traceable, auditable, and deterministic. No central orchestrator. Pure p2p choreography determined by the Wasm state machine. Think durable objects with formal states, but with no central infra. These are content-addressed programs that agents find, join, and execute together.

@ranjan3118 @huddle01com Chad @ranjan3118 ! Love your Clarity of thought and laser sharp communication every time I interact with you !

always awesome to keep testing waters. pitched @huddle01com at Supermoon Startup day in Dubai infront of a good panel of VCs and 100+ people and came 🥈 place 😄 nice to see people resonated w how big a problem current centralised bandwidth infrastructures are and how we at Huddle01 are building a demand first DePIN w our own app @MeetOnHuddle01 interestingly the 🥇 place startup also is building a super cool bandwidth DePIN network solving for egress costs kudos @Titannet_dao

@raulvk @ethereum Congrats @raulvk ! You are gonna kill it. Chad p2p stack coming up !

hey, i'm super enthusiastic about p2p, distsys, and networking. that's why i joined the Ethereum Foundation last week—to help accelerate @ethereum's p2p efforts 🚀 things like DAS, privacy-preserving protocols, peer anonymity, censorship resistance, light clients, decentralised rpcs, L2 interop, beam, etc. are hot topics 🔥 if you wanna bounce ideas around, reach out! ah, we'll probably be scaling this new team soon, so also hit me up if you wanna work on this stuff 🤝

@ranjan3118 @ranjan3118 The problem is that graduating from the right university or belonging to the right family is important in general back there if someone wants to be taken seriously. Therefore open source and open networks. Ukhaadlo jo ukhaad sako phir.

I might be guilty of this too subconsciously, but I believe most Indians who’ve reached a certain level of success or social standing don’t support other Indians who are trying to punch up. I’ve seen this visibly in the web3 startup ecosystem, especially in investor–founder relationships, but it runs even deeper. Culturally, this feels starkly different from how Chinese, Europeans, or Americans support each other. Can’t complain about others being biased against Indians when we ourselves are biased against each other. I don’t know how, but fragmented cultural trust and intra-community support needs to be addressed, and fast if we need to see more web3 projects from India at a global level.

@zeddotdev Hey @zeddotdec I am a long term Zed user and love it over al else. Please can I get access ? My github is aarshkshah1992

We've been working on an exciting new AI feature for the past few months and are now ready to expand our beta testing. Click the link below to join the beta waitlist! zed.dev/edit-predictio…

@zeddotdev Is there a way to automatically close a multibuffer if a reference search opens a multibuffer after you've navigated to the relevant symbol already ? I get too many open tabs without that which causes tab bloat that I have to then manually close :/

@rishdotblog @_skris @rishdotblog So I guess Cursor will wins for hacking on pre-existing mid-large sized codebases. I've only ever used Cursor but wondering if you had an opinion on how well Windsurf would work for non green field development on large pre-exsiting codebases.

@_skris Aye, much much better! Cursor seems to have released an improved Compose in beta earlier this week though (don't have access to it yet) – waiting to see how that pans out

I was surprised at how much better Cursor was compared to VS code, and am now surprised at how much better Windsurf is compared to Cursor Still a surprising amount of alpha in tinkering with different RAG/tool systems around frontier LLMs!

@ScottHConner @0xCygaar @ScottHConner I do worry about this eliminating junior devs. Why would anyone hire a grad fresh out of college now man ? Might as well get the LLM to write tests and implement data structures.

thought prompt engineering was a joke at first, but there really is an art in getting llms to give you good responses

Really enjoyed this talk on ways to find and measure PMF by the goat @j_schwartzz. "How disappointed will your users be if you take away your product from them" ?

@j_schwartzz @j_schwartzz Agree in principle. But this dosent work if you want to ship a complex protocol change or a high tech product. Imagine trying to build an LLM model or doing a complex cryptoecon change against a deadline. Move fast break things works better for user facing apps.

Anecdotally (haven’t done this myself), but I imagine when building a t shirt company from scratch, it’s better to design the shirt and set up your supply chain to print the t shirt after it’s ordered rather than invest money up front making hundreds of shirts. This protects the company from wasting money if no shirts are ordered Good software teams, esp with the help of AI, can deliver software products in a similar capacity. Invest as little as possible up front in a prototype, ship it, and iterate towards a polished product with real user feedback Many teams spend >6 months building to realize no one wants their product

@glifio is one of THE most underrated application in all of crypto. Really mean it. Solid product. Kudos to @j_schwartzz and team !

How it feels every time someone tweets that we’re underrated

@pooja_eth Brilliant and accurate !

My shitty, but I think true, drawing about life

@zeddotdev is craftsmanship. Blazing fast response times, intuitive UX,  a very no frills no magic LLM integration that exposes the entire LLM context to you and allows you to edit it just like you'd edit any piece of text and OPEN SOURCE. Rooting for them to succeed.

We’re talking part in @Filecoin’s FIL-RetroPGF Round 2 with @FILRetroPGF ! Check out our project to implement built-in Actor events in Filecoin here: #discussioncomment-10904759" target="_blank" rel="nofollow noopener">github.com/filecoin-proje…

We’re talking part in @Filecoin’s FIL-RetroPGF Round 2 with @FILRetroPGF ! Check out our project to improve the reliability and performance of the ETH RPC APIs here: #discussioncomment-10904518" target="_blank" rel="nofollow noopener">github.com/filecoin-proje…

We’re talking part in @Filecoin’s FIL-RetroPGF Round 2 with @FILRetroPGF ! Check out our project to add support for Legacy ETH transactions to Filecoin here: #discussioncomment-10904139" target="_blank" rel="nofollow noopener">github.com/filecoin-proje…

@j_schwartzz Will Win because they can take dem risks.

Whoever has less to lose