KSE @semanticbeeng

Shipping/bridging Engineering ⇆ Science #SoftwareArchitecture #FunctionalProgramming #MachineLearning #BigData #MachineLearningEngineering #CompilerDesign Europe Joined September 2011

Tweets

6K
Followers

769
Following

774
Likes

21K

Connor Davis @connordavis_ai

6 days ago

This MIT paper just broke my brain. Everyone keeps saying LLMs can't do real logical reasoning. Turns out we've just been teaching them wrong this whole time. These researchers built something called PDDL-INSTRUCT that actually teaches models to think through planning problems…

120 733 4K 283K 4K

Download Image

Andrea Barisani @AndreaBarisani

6 days ago

I am SSH'ing in my gaming PC bootloader to start Windows remotely. I am so grateful to the gVisor team for allowing me to put their pure Go TCP/IP stack on the bare metal.

1 1 5 2K 3

AlwaysLearning @WhisperingJack3

7 months ago

Google Cloud introduces quantum-safe digital signatures in KMS The cryptographic implementations will be open-source (via BoringCrypto and Tink libraries), maintaining transparency and allowing independent security audits. Google invites organizations to start testing and…

0 1 1 45 0

Quantitative Finance @QFinancePapers

a week ago

Adaptive and Regime-Aware RL for Portfolio Optimization. arxiv.org/abs/2509.14385

0 1 1 270 6

TACEO @TACEO_IO

a week ago

👨‍💻 Good news! Accessible for developers via : ✨ coCircom (docs.taceo.io/docs/co-circom/) ✨ coNoir (docs.taceo.io/docs/co-noir/) & our proof outsourcing network → TACEO:Proof docs.taceo.io/docs/taceo-pro… Access is no longer theoretical, it’s practical!

1 1 7 366 1

Carlos E. Perez @IntuitMachine

2 weeks ago

I just read a paper on "Drivel-ology" that completely shattered my perception of AI intelligence. My mind is still reeling. 🤯 Everyone says LLMs are getting scary good. They write code, pass medical exams, and generate stunning poetry. The conventional wisdom is that we're…

8 7 30 4K 30

spacy @dosco

3 months ago

the axflow idea is turning out soo good. get all the goodness of dspy and ax, streaming, tracing, error correction, signatures, multi-step function calling, etc across your entire llm program and the whole thing is optimizable as well.

3 1 10 675 3

Download Image

Eugene Yan @eugeneyan

2 weeks ago

Demo of the Qwen3-recommender hybrid returning both semantic IDs and natural language! • steering recs via natural language • explaining the recommendation • naming the bundle of recommendations • multi-turn conversation to get recs watch till the end for the bloopers lol

6 18 128 12K 115

Download Video

Quantitative Finance @QFinancePapers

2 weeks ago

ProteuS: A Generative Approach for Simulating Concept Drift in Financial Markets. arxiv.org/abs/2509.11844

0 1 3 346 2

cognee @cognee_

3 weeks ago

September @cognee_: 10+ launches, 4 case studies, 7 events, 8 collabs. Agents fail without memory. Good AI memory needs structure, relevance weights, temporality, evals + auto-opt, scale & governance. That’s cognee. We'll build, integrate, prove through the month. Join us!

0 2 7 290 2

Download Image

Vasilije @tricalt

2 weeks ago

We mapped the invisible threads of a city’s student community. @knowunity brought us a dataset of 40,000 learners - grades, school types, usage timing, and anonymized IP histories (privacy-first). The goal: move beyond content into connections that create real collaboration…

1 3 9 355 6

Download Image

🔥 Kareem Carr PhD 🔥 @kareem_carr

2 years ago

Causal Inference is a revolution in how we do science. Here's why: 1. Even more than statistics, causal Inference is the true mathematical language of science The primary contribution of causal inference is linguistic. Causal inference is a pair of mathematical languages. The…

28 156 832 249K 807

Download Image

Rose Yu @yuqirose

a year ago

How to discover multiple causal graphs from heterogeneous time series? Check out our #icml2024 paper on ``Discovering Mixtures of Structural Causal Models from Time Series Data''! Paper: arxiv.org/abs/2310.06312 Code: github.com/Rose-STL-Lab/M… (1/3)

1 25 161 18K 86

Download Image

Archetype @archetypevc

2 months ago

x.com/i/article/1948…

4 5 28 3K 19

TACEO @TACEO_IO

2 months ago

We built coNoir to make Private Shared State usable for builders in the @NoirLang and @aztecnetwork. It’s still early, and in active development, but the first wave of apps is already emerging. 🔧 Docs: docs.taceo.io/docs/co-noir/

1 2 19 1K 1

Noir @NoirLang

2 months ago

Zero-knowledge proofs allow for ownership of private date, but they're inherently stateless. That makes real world apps difficult without compromising privacy. @luhelminger from @TACEO_IO breaks down how Private Shared State fills this gap. youtube.com/watch?v=HPDWav…

4 6 31 4K 4

Noir @NoirLang

4 months ago

The NoirCon 2 speaker list just dropped. 👽 Kicking us off strong is @TACEO_IO CEO, @luhelminger with a keynote: Private Shared State: The Missing Layer. He’ll explore how coSNARKs enable collaborative computation on encrypted data with verifiable results, unlocking a new…

8 5 57 6K 4

Download Image

Matthew Green is on BlueSky @matthew_d_green

8 months ago

I wrote a post about how AI will get along with end-to-end encryption. TL;DR: maybe not so well! blog.cryptographyengineering.com/2025/01/17/let…

10 77 252 30K 115

Tinfoil @TinfoilAI

7 months ago

Data security remains the #1 barrier to enterprise AI adoption. At Tinfoil, we're solving this with secure enclaves that provide end-to-end confidentiality for AI applications—letting companies verify nobody else can access their data. Learn more: tinfoil.sh/blog/2025-02-2…

1 2 6 320 0

PCMag @PCMag

4 weeks ago

Anthropic once deleted Claude chats after 30 days — and didn’t train on them. Now, it wants to keep your data for 5 years — and use it to train AI, by default. A pop-up is coming asking you to agree by Sept. 28. Opt-out available. 🔗 pcmag.com/news/anthropic…