🚨 THE AI COST CRISIS HAS STARTED. Microsoft reportedly told engineers to stop using Claude because AI bills were exploding, while Uber says its entire yearly AI budget was already destroyed by April.

Just found out that Berkeley course staff are writing hooks inside course repos so if a student opens an assignment in Claude Code or Cursor the agent will automatically ping the staff 😵‍💫 well played

shooting my shot: I'm dating again! I'm a 31-year old eng in SF, mostly nerdy (into local politics, the economist, a24, Ellsworth Kelly) also reasonably outdoorsy (running, skiing) and I cook you should be a 25-35 guy who wants kids, is intellectually curious, and lacks chest hair

Happy Friday! We've reset everyone's 5-hour and weekly rate limits.

Can AI write EVM bytecode a Lean proof of solvency under arbitrary reentrancy, bypassing the compiler entirely? Yes! In this experiment we create 86 bytes of WETH bytecode plus a sorry-free Lean solvency theorem 👇 (thread link below)

Some time ago, I had the idea to port NVIDIA Physical AI stack to AMD. The motivation was to improve hardware diversity and enable world models and VLAs to run beyond a single ecosystem. We started with NVIDIA Cosmos Predict 2.5-2B. Porting wasn’t trivial: these models are deeply optimized for NVIDIA’s stack. We used this as an opportunity to apply our ROCm kernels. The results were surprising: Both encode and diffusion run faster on AMD Instinct MI300X vs. NVIDIA H200 (FA3) and we still saw significant headroom for further optimization. Quality is unchanged across modalities (validated with WorldJen) To be clear, this is no luck. We have deep experience with diffusion models and AMD GPUs. But this just gives us a good opportunity to get closer to a true hardware-to-hardware comparison, as we work with less software abstractions than usual. Just to give an example, on AMD, memory instructions are async with a hardware queue of ordered pending instructions, enabling concurrent load/store with compute without warp specialization. Bottom line: there are real architectural advantages on AMD, if you take the time to work with the hardware. Note, we did tradeoff ~20% higher memory usage, That being said, AMD has more to give to begin with :) in the coming weeks: AMD versions of Cosmos Transfer and GR00T, an even faster version of Cosmos Predict, and open-sourcing an attention kernel faster than AITER v3 (which is closed-source for some reason? cc: @AnushElangovan )

Great visual from @Ethan_Heilman on why taproot addresses are quantum vulnerable at the #MITBitcoinExpo @MITBitcoinClub

Quantum-Safe Bitcoin Transactions Without Softforks github.com/avihu28/Quantum-S…

be me try to write short zk on bitcoin piece think, "let's peep @AlpenLabs code base and use it as an example" realize it's a godly repository study and start writing example get scared. might be wrong message @niftynei "i need a few more days" take xanax lock back in

Live version: trustmebro.starkwarebitcoin.… Github repo: github.com/AbdelStark/trustm… Few notes: - current version only has transaction verification and not yet block header verification (will be included in next version shortly) - this won't work for recent blocks: there is some delay for the proof generation and we have a notion of minimum confirmation time / size of chain to prove blocks - the STARK proof verification runs in the browser using a WASM STWO verifier

I just got a 100KvB transaction mined, full of OP_Returns containing the NASTIEST malware and viruses known to mankind, for Dos, Windows, Linux, Android, and MacOS!!! WATCH AS EVERY BITCOIN NODE DIES!!! @GrassFedBitcoin SAID IT WOULD BE THE APOCALYPSE!!! mempool.space/tx/85f1bf57386…

NEW WORLD RECORD. WE CAN PROVE THE WHOLE BITCOIN BLOCK HEADER CHAIN IN 6.5 HOURS. FOR $13.5. YES, WE HAVE PROVEN 16 YEARS OF BLOCKCHAIN HISTORY OF A $2 TRILLION ASSET FOR LESS THAN A BIG MAC COMBO MENU. STWO, FASTEST PROVER IN THE WORLD, BACKBONE OF STARKNET.

Since I’m turning 26 and haven’t been in a relationship for 2 years, I’m now seriously looking for a boyfriend Not because I’m lonely but feel there’s so much joy I can share in my life. We can go to beach together, snowboarding together, or just talk laugh and couch potato together Dating me pros: fun & sarcastic, loyal no drama, very much caring, not retarded (sometimes Hope you are: 1’ into BTC 2’ into Asian, at least sometimes 3’ shave beards (is a must 4’ kind 5’ know the difference between Fiamma bridge (BitVM2) and multisig bitcoin bridge That’s all 1BTC referral fee if succeed Dm closed and long term valid :) (You can networking your way to reach me if you are the right one.)

I got tenure at UCSB! 🎉 12 years ago, I knew almost nothing about PL. Thank my superstar advisor, Isil, for teaching me how to do real research (and sat through those 10-hour pre-PLDI meetings from sunrise to sunset). We were the best partners on almost everything. I was also extremely lucky to be mentored by THE BEST PL folks on this planet, Alex Aiken, Ras Bodik, and Emina Torlak, who shaped my research taste. To my amazing students—Yanju, Junrui, Hongbo, @trackoor, Bryan, Nicholas, and many others—you are the best part of this job. To my collaborators—Osbert Bastani, Ruben Martins, Yuan Tian, Chenglong Wang, Yuepeng Wang, Xinyu Wang, @lukejdpearson (also my investor, lol)—thank you for making research a true joy. To my UCSB colleagues and the anonymous letter writers—thank you for the incredible words and believing in me. And to my wife—thank you for standing by me through those tough years and every crazy decision along the way. I couldn’t have done this without all of you. Here’s to the next chapter. 🚀

Contributing to Citrea’s BitVM TSC 🫡 Should I destroy my computer now? 👀👀👀