871
Let's be honest, Generative AI isn't going all that well
A sampling of recent news
Top posts
Every blog post that took off on Hacker News, ranked by the votes of its best submission. 1,521 posts.
872
Discovering Cryptographic Weaknesses with Claude
Anthropic researchers find weaknesses in cryptographic algorithms with Claude Mythos Preview
873
Better Models: Worse Tools
About an aggravating tool-calling regression in newer Claude models.
874
Opening up 'Zero-Knowledge Proof' technology to promote privacy in age assurance
Today, we open sourced our Zero-Knowledge Proof (ZKP) libraries, fulfilling a promise and building on our partnership with Sparkasse to support EU age assurance.
875
AI 2040 and the cult of intelligence
I used to be one of these people. I read Yudkowsky and was like, OMG recursive self improvement hard takeoff AI is coming. Then I joined the real world and actually tried to do things. At comma, we ship a hardware product of similar complexity to a cell phone, and it’s really hard. Reality has lots of finicky details…
876
AI may be making us think and write more alike
Large language models may be standardizing human expression and subtly influencing how we think, says study led by USC Dornsife researcher
877
Once Claude can measure something, it can make it faster
Inside our performance sprint: the benchmarks Claude built, the loop each Slack thread ran, and the guardrails that let us ship 3,000 changes safely.
878
Step 3.5 Flash – Open-source foundation model, supports deep reasoning at speed
Scores represent the mean of the following eight benchmarks listed below, excluding xbench-DeepSearch. The Step 3.5 Flash score is derived under standard settings (i.e., $w/o$ Parallel Thinking).
879
Trinity large: An open 400B sparse MoE model
A deep dive into the open weight Trinity Large model, covering architecture, sparsity, training at scale, and why we shipped Preview, Base, and TrueBase checkpoints.
880
AI-assisted cognition endangers human development?
Does AI-assisted cognition threaten human development? Explore the risks of AI-assisted thinking and learn strategies to use AI tools without freezing your critical thinking.
881
Use boring languages with LLMs
Why coding agents perform better in cohesive ecosystems like Go than fragmented ones like JavaScript or Python
882
The ways we contain Claude across products
As agents grow more capable, so does their potential blast radius. Here's what we learned building containment for claude.ai, Claude Code, and Cowork.
883
Munich 1991: The Roots of the Current AI Boom
1991: T and P of ChatGPT, distillation, deep residual learning, LSTM, GAN
884
You only need the frontier model for one single edit
/prewalk hands the cheap model a trajectory, not a plan, retaining 92–97% of frontier pass rate at roughly half the cost on SWE-Bench Pro.
885
Training AI to Paint with Code
I'm a designer and creative technologist based in Brooklyn, NY.
886
Humanising LLM Outputs Is Dumb
The largest tell for me to tell where culture and sentiment is shifting for AI tools is usually X, viral GitHub repositories and Hacker News. One of these tells I’ve been seeing a lot lately is skills like I have ADHD and Agents.md instructions such as giving outputs in only ASD-STE100 Simplified Technical English.
887
Using Opus 5.5 to discover a new eyewitness record of the dodo
Frontier models can now produce novel historical knowledge, but in a really weird way
888
Anthropic's growth just flattened
Look at the monthly numbers...
889
Access to frontier AI will soon be limited by economic and security constraints
Soon, access to frontier AI will be scarce and selective
890
891
Using LLM for confidence scoring is a bad idea
LLM-based agents can do a lot of tasks. Or, at least, they can appear to do them. For example, they write code, summarize papers, plan trips, and fill in forms. If you ask one for a probability or a confidence score, it will give you one of those too. I have seen many agent designers use LLMs to generate confidence…
892
Mxc: Microsoft Execution Containers version 1.0.0
Agents are unlocking enormous productivity gains for customers, but their ability to work across files, networks, and applications can introduce new security risks. This can leave customers feeling like they only have two choices: give agents unrestr
893
If DSPy is so great, why isn't anyone using it?
Any sufficiently complicated AI system contains an ad hoc, informally-specified, bug-ridden implementation of half of DSPy.
894
Claude Platform on AWS
The Claude Platform on AWS is now generally available, offering a new way for AWS customers to access the full set of Claude platform features with AWS authentication, billing, and commitment retirement.
895
FLUX.2 [Klein]: Towards Interactive Visual Intelligence
Introducing FLUX.2 [klein], our fastest image models. Unified generation and editing in under a second, on hardware you already own. Apache 2.0 on 4B.
896
Fable Took My Job
Fable is cool yo, it takes jobs yo, - Dario "steve" Altman
897
Claude now creates interactive charts, diagrams and visualizations
Ask Claude to explain a concept or analyze your data, and it can respond with interactive charts, diagrams, and visualizations — rendered inline as part of the conversation.
898
The Q, K, V Matrices
At the core of the attention mechanism in LLMs are three matrices: Query, Key, and Value. These matrices are how transformers actually pay attention to different parts of the input. In this write-up, we will go through the construction of these matrices from the ground up.
899
Fable and the end of the free lunch
Fable’s price ended the free lunch of ever-cheaper frontier models. Now we have to think about which model gets which work.
900
Don't Trust the Salt: AI Summarization, Multilingual Safety, and LLM Guardrails
“The devil is in the details,” they say.