Luke Wang

@raphaelschaad So true, it’s almost a double reinforcing loop, I feel a bit of field experience managing/orchestrating human teams goes a long way improving AI usage too

Ironically, instructing AI has made me a more capable orchestrator of … people. Before, I felt irked sharing memories, stating preferences, and setting boundaries. Now, I send my accountant an .md file with all 2025 context, and they’re off to the races.

@AbdelStark Amazing

Zm. 55 projects that have submitted for our starknet track. That's it, that's the tweet! Ztarknet, Zypherpunks hackathon, freedom tech builders.

@raphaelschaad British kitchens with the laundry under sink?

Microwave over stove is awful design, change my mind?

@0xfishylosopher Reminded of first principles is golden

i love talking with students just getting into crypto - its a massively underrated opp - not just for finding the next generation of talent, but also as a way to test and pitch your own understanding of crypto as a space, crypto often gets lost in the internal subculture of CT, swallowing itself in nihilism rather than looking at the broader picture. when you talk to students unfamiliar with ct jargon, you're forced to trace back to first principles - "global permissionless money", "buying stocks anywhere, anytime", "a world computer that proves its code execution", "predicting any real world event".... and so on. add to this the unique lifestyle of the space - the ability to travel the world with your friends, the ability to make a real impact straight out of college, an alternate way to make money etc.... you realize that crypto IS in fact competitive with things like AI, quant, banking/consulting, simply because the opportunity per capita is sooo much greater than any other industry. the "opportunity per capita" of crypto is fundamentally what made me take a leap of faith into the space, and what i've been constantly yapping to @cynthwangg and other folks at stanford in my cryptopilling role in the past 3 years. imo crypto moves at maybe 4x the speed of any other industry. but sometimes its worth to breathe, take a step back, and look at how far we've come along in a few short years. wagmi.

Got a DM three days ago from someone using Nora to prototype quantum-resistant protocols — they’d already maxed out our Master tier. We shipped extra-credit top-ups ASAP. Crypto might be in a reset phase, but the level of stuff people are building right now is unreal

Thanks for inviting, excited to 100× the on-chain dev experience for more developers with AI

Appreciate the invite to mentor at Binance Chain’s hackathon — excited to support the next wave of crypto builders as the AI tools we’re building at @mynoraai unlock a new paradigm of Web3 development.

@sunbh_eth wen

just arrived buenos aires. wen asado?

@0xfishylosopher lol, can't say I resisted laughing

@tkstanczak Bonkers idea if this was available for free and cost of a laptop + internet instead of 40k a year

Would be great to create an open source education curriculum: cbsnews.com/video/inside-t…

@chatgpt21 No, outside of a few startups that focus on this, this should create way more opportunities by making it 10x easier to build agents for non-devs and casual creators.

OpenAI just killed 1,000,000,000,000,000 Startups 🤯

@connordavis_ai There's actually a field/topic called MARL (multi-agent reinforcement learning) that studies just this, but it rarely works for anything beyond 3-5 agents. It's an amazing field. You get to see AI learn key collab concepts from zero, but so much more progress needed

Multi-agent AI is a $50B lie. 99% of "multi-agent" systems are just single agents with fancy marketing. I just read the paper that exposes what real multi-agent intelligence actually looks like. Most people think multi-agent AI is just "multiple ChatGPTs in a room. That's like saying a surgical team is just "multiple people with knives." The real story is way deeper. Task allocation is completely broken. Current systems are basically throwing darts at a board. Give the math problem to whoever's free. Ask the creative agent to debug code. It's chaos disguised as intelligence. Real multi-agent systems need dynamic specialization. Not just "Agent 1 does X, Agent 2 does Y" but context-aware matching based on capability, workload, and past performance. The memory problem is insane. Single agents just track conversations. Multi-agent systems need five different memory types: short-term task state, long-term expertise, episodic collaboration history, consensus knowledge, and hierarchical access control. Most current systems give every agent amnesia between tasks. Context management is where everything breaks. Each agent needs to track three layers simultaneously: the big picture mission, their specific piece, and what everyone else is doing. Fail at any layer and the whole system becomes expensive nonsense. Game theory matters more than code. When agents debate or negotiate, you're not optimizing for "correctness." You're finding equilibrium states. The research shows Stackelberg dynamics work better than Nash equilibrium for most real tasks. Nobody talks about this because it's not as sexy as "look, the robots are talking." The applications they outline are wild. Agents that negotiate smart contracts autonomously. Fraud detection where different specialists hunt different attack patterns. Consensus mechanisms that actually think through decisions. We're not building better chatbots. We're building the foundation for autonomous economic systems. The gap between current "multi-agent" demos and actual multi-agent intelligence is massive. Real systems will have specialized roles, shared memory architectures, and game-theoretic coordination. They'll solve problems no individual agent can handle. Same principle that makes human teams work. Just faster, and at scale. Most of what people call "multi-agent" today is just single agents with fancy prompting. The companies that figure out real multi-agent coordination first will have a 10x advantage. Everyone else is building expensive theater.

@gregisenberg Too true

it’s 100x harder to go from $0 to $1k MRR than to go from $100k to $1M MRR

@EliBenSasson bullish on StarkNet

Bullish on Ethereum, bullish on Bitcoin, bullish on Starknet, bullish on ZK.

@mxstbr @Shopify @ryolu_ This is why the impact and disruption of AI code is far beyond just helping devs code slightly faster and sell credits for profit. The fundamental way we work and collaborate will change

Design → dev handoffs are a thing of the past. At @Shopify, our designers vibe code prototypes of their ideas in playground/"baby" versions of the product. (h/t @ryolu_) They then work with devs to implement the prototypes for real. There's no "handoff" along the way. It's one continuous stream of work and collaboration. One team.

@raphaelschaad Oh, this is a good one

Try Nora

“Product” has sadly become an anti-principle in Web3 — something veterans have come to instinctively avoid. But what happens when AI comes? When build → ship takes hours, not months? The constraints change. The game changes.

50 million H100s would use as much power as the UK or France — about 50 nuclear plants. What hits first: • Elon builds 100× better AI chips, Tesla > Nvidia • Or we get Nuclear Elon

@abdelstark Starknet will melt faces

👀 Malachite, Starknet's consensus implementation ready to test integration with full nodes and sequencers. Be the decentralisooor brother. TL;DR: Starknet will melt faces. #swmf