Ben Livshits @convoluted_code

A technology executive, Computer scientist, Professor. A decade at Microsoft Research, executive roles at Brave, ZKSync, Eclipse. Worldwide Joined April 2015

Tweets

1K
Followers

9K
Following

4K
Likes

1K

Wei Dai @_weidai

3 days ago

This is cool--Karpathy's autoresearch idea applied to a real load-bearing problem: minimizing the size of a quantum circuit breaking DL over secp256k1. Recall that the smallest quantum circuit currently known (from a group of researchers from Google and more) was not released publicly, and the authors only proved knowledge of a quantum circuit via zk. This is basically completely the opposite: the smallest quantum circuit will be publicly available, and anyone or any agent/AI in the world can contribute to it! How the lower bound develops for the next few weeks will be very interesting to watch. My bet is on the move to post-quantum cryptography moving to an even more accelerated timeline due to this project.

gautham @bbuddha_xyz

3 days ago

I beat one of the best published quantum circuits for breaking Bitcoin. And I have no formal training in quantum cryptography. Using just AI agents, I improved it by ~2x. But I haven’t beat Google’s best classified circuit yet. So, today I'm launching ecdsa(dot)fail -- an open

53 55 422 165K 171

6 21 103 17K 50

View Details

Ben Livshits @convoluted_code

a week ago

So what is the minimal amount of human supervision needed for surpass agreed-upon quality gates?

Yoav @YoavCodes

2 weeks ago

@clirocks @jarredsumner We don’t have to wait for the law or expect unserious people to suddenly behave seriously just because there might be litigation. You can’t change the physics of obvious outcomes from shipping unmitigated slop. It’ll resolve itself because slop products will be evidently worse.

0 0 8 1K 0

0 0 0 265 0

View Details

Ben Livshits @convoluted_code

2 weeks ago

I agree with a lot here but I think the infrastructure vs. business point is overdone — it’s possible to find the right middle ground as has been done elsewhere for decades

Simon Dedic @sjdedic

2 weeks ago

I've been an Ethereum supporter since the early days and I still believe it's one of the most important pieces of infrastructure in crypto. The tech is great, my conviction hasn't changed in that respect. But watching 9 senior researchers and key operators leave the Ethereum

147 39 592 174K 120

1 0 1 365 0

View Details

Prakash @8teAPi

2 weeks ago

So Jane Street is going public because obviously they see the future where the model labs compete directly with them in the market. The strategic decision is therefore to become a a specialized infrastructure harness for a future frontier model. Tellingly they point out that the latency constraints mean there is no time for inference at the GPU layer, or agentic tool use at the CPU layer, only reflexive heuristics at the FPGA layer. @yminsky is trying to fend off future model lab competition by making Jane Street indispensable to a future AGI. interesting strategy

31 62 2K 155K 928

View Details

Rohit Chauhan @degenrsc

2 weeks ago

Beautiful piece on investing in the AI trade x.com/ramahluwalia/s… tl;dr - AI investing or sector as a whole is now operating the power law game ie one or two players own a LOT, and rest go to zero (expensive science projects with a pitch deck, unicorn valuation (for now), and zero revenues. Meanwhile Anthropic kills it and owns everything - AI isn’t flattening the economy (equal access or upside), rather steppening it (winners takes all) - average isn’t safe. If you’re thinking you can spread your capital and win, you’re in for a violent surprise. Pick the winners, cut the losers and do it now Wonderful read, worth your time folks

Ram Ahluwalia CFA, Lumida @ramahluwalia

2 weeks ago

FEAST OR FAMINE We have a K Shaped economy already. AI is a knife that accelerates that. @GavinSBaker makes a great point. Anthropic has added tens of billions of ARR in a matter of months. That scale that took elite software companies like $SNOW, $PLTR, and Databricks ten

4 0 30 9K 7

1 1 1 732 1

View Details

Ben Livshits @convoluted_code

2 weeks ago

Worth reading

Laura Shin @laurashin

2 weeks ago

I think Ethereum’s original sin was not considering tokenomics with every move it made from Dencun on. The ultrasound money thesis was a good one and with Dencun (or the L2 roadmap generally) they should have stopped to say that this was going to hurt the ultrasound money thesis

139 51 597 293K 151

0 0 0 298 0

View Details

Bojan Tunguz @tunguz

2 weeks ago

Try to build as much code as possible over the next few months. The prices you are seeing now for AI will probably not last too long.

Hedgie @HedgieMarkets

2 weeks ago

🦔Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effectively infinite cloud resources. Uber's CTO sent an internal memo warning the company burned through its entire 2026 AI budget in just

1K 4K 20K 8.3M 12K

152 300 4K 1.2M 2K

View Details

Ben Livshits @convoluted_code

3 weeks ago

While many runtime enforcement startups are emerging, it remains to be seen which combination of techniques will gain long-term traction.

0 0 1 73 0

View Details

Ben Livshits @convoluted_code

3 weeks ago

I suggested constrained decoding, but there are some attractive options, where code is generated directly in Lean or a combination of Lean and Rust is generated side-by-side.

1 0 1 86 0

View Details

Ben Livshits @convoluted_code

3 weeks ago

A while back, I argued that we must move beyond post-hoc vulnerability discovery. With agentic AI making automated exploitation a reality, proactive defense is no longer optional. 🧵 arxiv.org/abs/2602.08422

3 0 4 356 0

View Details

Kyle Jeong @kylejeong

4 weeks ago

x.com/i/article/2051…

26 87 1K 952K 3K

View Details

Wei Dai @_weidai

a month ago

A highlight of @1kxnetwork investments within our thesis on threat-resistant & compliant onchain privacy. - @0xMiden: chain designed from the ground up for programmable privacy (ZK) - @zksync: private & customizable prividiums (ZK) - @inconetwork: user-friendly privacy layer for existing chains (TEEs) - @SeismicSys: privacy-enabled EVM-based fintech L1 (TEEs) - @ligero_inc: private account layer for all chains, custom-built for businesses (ZK) - @0xPredicate: programmable policy infra, for privacy protocols, defi, & beyond - @fiber_evm: private EVM wallet infra w/ a slick mobile app (ZK) Here's the exciting bit: (at least) five of the above projects are about to go live this year! 2026 is the year for onchain privacy. ----- At 1kx, we put our money where our mouth is. We develop theses and partner with the best founders to realize the shared vision.

Wei Dai @_weidai

a month ago

Onchain finance needs threat-resistant privacy → Real-world & institutional finance cannot move onchain without privacy → To prevent misuse (e.g. laundering of hacked funds), the only viable solution is to build threat-resistant privacy More in my op-ed in Forbes 👇

7 12 72 40K 18

13 23 131 22K 51

View Details

Adam Cochran (adamscochran.eth) @adamscochran

4 weeks ago

The day after the CEO lays off a ton of staff and says: “Non-technical teams are now pushing code to production with AI” @coinbase has a major outage on their trading engine, and even their status page doesn’t work. 😂

Steven @Dogetoshi

4 weeks ago

Their status page is also down 😭

50 67 1K 2.7M 105

331 2K 15K 986K 999

View Details

Gary Marcus @GaryMarcus

a month ago

Some things never change. If you don’t understand this one, you don’t understand what’s happening AI. Marcus, 1998: neural nets have trouble generalizing far beyond the data. Marcus, 2001, 2012, 2019, 2022, etc: neural nets have trouble generalizing far beyond the data. Apple, 2025: neural nets have trouble generalizing far beyond the data. Meta/Stanford/Harvard, 2026: neural nets have trouble generalizing far beyond the data.

Deedy @deedydas

a month ago

The creators of SWE-Bench just dropped a really simple new benchmark every LLM gets 0% on. ProgramBench asks: can models recreate real executable programs (ffmpeg, SQLite, ripgrep) from scratch with no internet? We are far from saturated on model quality.