My first experiment with LLM-assisted coding: an editor for writing sequence of hybrids. There is no formal verification; it's just display logic for "pen and paper" security proofs, meant to help the author reason about the sequence of hybrids "horizontally" and "vertically".
since a lot of folks asked me, this is the research agenda I have in mind for **generative cryptography**
> "what can we unlock if AI could write cryptography?"
1) ai agents can engage in cryptographic interactions
2) autoresearch can improve protocols
3) solve open problems
Yes. $intc 14A hss secured multiple foundry customers. They arent announcing those yet to save the clients backlash from $tsm. Tbh the chip demand supply constraint is so huge that both intc and tsm can have significant market shares...in the same space.
AI agents running on your machine have shell access, file access, API access and zero runtime policy enforcement.
Built CATP: open-source enforcement layer for Claude Code. Define a TOML policy. Get real-time blocking + a tamper-evident audit log.
https://t.co/eYjeiXWxXU
no one is actually building a Github replacement
cursor's origin launch just confirmed this - they are only building a replacement for the _paid_ version of Github, because that's where the money is
if everyone goes after the paid source control business and take that away from Github, Github will not be able to sustain the open source ecosystem
i hope we figure out a viable alternative before that happens..
Biggest notable adds to $INTC in the Q2 2026 13Fs:
1. SoftBank Group: Largest overall holder with 86.96M shares. (Japanese tech conglomerate led by Masa Son).
2. Tiger Global Management: More than doubled to around 4.3M shares (~$594M), now ~2.5% of portfolio. ($50B+ New York growth hedge fund run by Chase Coleman)
3. Assenagon Asset Management: +45.5% to 16.15M shares ($2.3B). (Independent Luxembourg-based European asset manager, ~€83B AUM)
4. Nordea Investment Management: Doubled to 4.76M shares ($654M). This is the asset management arm of Nordea Bank, the largest Nordic bank, ~€280-300B+ AUM)
5. Robeco Institutional: +74% to 3.75M shares ($523M). (Dutch asset manager owned by Japan’s ORIX Corporation, ~€200-230B AUM)
6. Daiwa Securities Group: Nearly doubled (+99%) to 3.82M shares ($533M).
Intel leads all stocks in number of new positions so far (~40 funds initiating), based on 13F filings.
I've heard of math PhD students losing motivation because AI is solving everything. This is nonsense. If I look at my PhD field - computational complexity theory - the big questions are all just as open as they've been for decades. We barely know anything. We can't prove things that seem trivial and obvious at first glance:
As one example, we can't prove that a poly time algorithm can't always be simulated in logarithmic space!
Like in the metaphore with two men running away from a bear - it's easier today to have an edge over your competitors. People are so distracted, and most people don't even think that's a problem because they believe AI can do their thinking for them better than they can, so why go to the trouble of focusing?
Carve a chunk of the day where you're doing things old school - just you a pen and paper and maybe a book. No internet access for a continuous chunk of time. I'm solving (trying :) ) exercises from silverman EC book at the moment.
PPl who will preserve the ability to think humans used to have till fairly recently will have a huge advantage - like the ppl who invested in crypto in early 2010's.
That's at least my (contrarian? obvious?) opinion.
Day 3 lectures:
@sourav1547 on Distributed Key Generation: Definitions, Constructions, and Applications:
https://t.co/WiOVXIAkjS
@guruvamsip on Encrypted Mempools and Batched Threshold Encryption:
https://t.co/vLfq6u8flJ
@alberto_sonnino on DAG-Based Consensus: Introduction and Practical Aspects:
https://t.co/LnITagYU2b
Post Quantum Crypto for Ethereum.
Justin Drake and the EF crypto team are pushing the frontier of hash based cryptography for PQ Ethereum. Among the many considerations, hash based systems have the least known attacks in the post Quantum crypto universe, particularly if you are using EXISTING hash schemes, this minimizes the time to deploy.
There is a recent attack claimed on the other front runner (Lattice PQ) by renowned Quantum cryptographer Simon (preliminary draft) - not yet verified - https://t.co/pzjpyUSr4K
It is our pleasure to contribute to accelerating research for Post Quantum Ethereum.
We started this with ecdsa (dot) fail building better Quantum attacks on ecdsa. Then with snark (dot) fast we worked with EF and Succinct to accelerate prover speeds 2.5x for hash based PQ cryptography.
We are looking forward to do much more with our Open Frontier Research platform https://t.co/fa48nQizUP
Come accelerate the frontier of cryptography with us!
A new fast binary-field proving system is out: Flock by @benediktbuenz, @ronrothblum and @kleptographic.
We'll be spending the next couple of weeks exploring whether it could be a good fit for proving post-quantum signature aggregation for Ethereum. 🤞
https://t.co/i7O0dsAmDv