Liquid AI caught its model cheating on coding tests — then it recognized the sandbox
Liquid AI’s new reasoning model was built to solve coding problems, but during training it kept combing Git history, GitHub, The post Liquid A…
Последние новости Web от Tech News
Все темы
AI
agents
ai
api
architecture
automation
backend
beginners
career
database
devchallenge
devops
discuss
javascript
linux
llm
machinelearning
mcp
opensource
performance
productivity
programming
python
react
security
showdev
testing
tutorial
typescript
webdev
Anthropic built its own git replacement for AI-scale coding. But you can’t have it.
This week, Anthropic revealed it built an in-house version control system to handle exploding code merges. It says Claude was both part Th…
Agentic AI doesn’t have a model problem. It has a retrieval problem.
On BrowseComp-Plus, GPT-4.1 answered 93.49% of the questions correctly when it was given the supporting documents directly. When an agent The post A…
Linus Torvalds let Claude Code fix the Linux page cache. Then reviewers found a race.
Linux creator and longtime lead maintainer Linus Torvalds turned to Anthropic’s Claude Code to tackle a kernel performance problem that The po…
You can outgrow vanilla Postgres without abandoning Postgres
For many data aficionados, Postgres is the way to store and manage relational data. It’s efficient, it’s straightforward, and it The pos…
Most AI incident tools still leave a key decision to humans
An AI SRE agent can help troubleshoot an incident. But can you trust it to decide on its own when The post Most AI incident tools still leave a key…
The AI safety check that runs on a laptop and nearly matched a 35B model
Guardrail selection has typically meant choosing between a purpose-built classifier and an LLM acting as a judge. Decision models such The post The…
Bit Cloud’s next chapter starts after the AI builds your app
For many developers, building an application from an AI prompt is already part of the workday. Turning that first version The post Bit Cloud’s next…
OpenAI’s Dots boundary problem rate doubled in longer tests
OpenAI’s new Dots are built to keep working after you step away. Launched at DevDay on Tuesday, the always-on agents The post OpenAI’s Dots bo…
OpenAI answers TypeSafe’s Jev with a Decision API built on Luna
With the sudden rise of TypeSafe’s Jev, decision models have become incredibly popular, so it’s not surprising that OpenAI is The post O…
OpenAI blocked its agent’s web access. Then it tunneled out through DNS.
Last week, OpenAI published new reports detailing misalignment issues with its agents. This time, two internal models found ways around The post Ope…
Think you need a better reranker? Check your retrieval first.
You built a search system using the latest, greatest models, and maybe you were thrilled with the outcome. But what The post Think you need a better…
Performance engineering from kernel analysis to AI: Adrian Cockcroft’s take
Over its five-year history, P99 CONF has hosted quite a few speakers who’ve offered pointed takedowns of the namesake metric. The post Perform…
Query decomposition doesn’t fix context starvation — it just moves it
I have a small example that would best communicate the message I am trying to convey: say you built a The post Query decomposition doesn’t fix…
Vulnerability alert fatigue nearly swamped WHOOP. But its fix still keeps a human in charge.
With often hundreds of thousands of alerts a day, many tech organizations are buried in vulnerabilities and worn down by The post Vulnerability aler…
AI spending can run negative. Qodo’s CEO built an ROI equation to fix it.
Flush with the proceeds of a $70 million Series B raised earlier this year, you might expect Qodo to spend The post AI spending can run negative. Qo…
Grok Build vs. Claude Code: I tested which one has the better memory
On September 16, xAI announced memory in Grok Build, its terminal coding agent. The pitch was that Grok “keeps notes The post Grok Build vs. Claude…
This week’s news from Zed, Anthropic, and OpenRouter shows why better harnesses matter more than better models
Our five most-read stories this week covered code collaboration, a model router, a UI change, a benchmark, and a caching The post This week’s news f…
“Be transparent only if asked”: OpenAI’s models learned to leave notes for their future selves
OpenAI revealed Wednesday evening that some GPT-5.6 Sol model instances, during reinforcement learning (RL) training, wrote instructions to conceal…
Why human oversight is shifting from writing code to defining requirements
This walks through the pipeline our agents operate inside—from a recorded scoping meeting through unit specs, spec review, generated code, The…
Bolt is giving developers 50x more compute. But there’s a catch.
Bolt.new, StackBlitz’s browser-based AI development platform, is testing a new trade with developers: more coding-model usage in exchange for…
Chip Huyen explains how to cut inference costs without new hardware
Last October, the P99 conference — the online gathering for developers focused on high-performance, low-latency applications — featured…
OpenAI split a voice model’s brain. Then one team deleted 23,000 lines of code.
Building an AI voice agent has always been clunkier than it seems. Most voice agents are really a chain of The post OpenAI split a voice model’…
Shopify spent years on React Native — then rebuilt everything in 12 weeks
In 2020, Shopify made a bet that resonated across the developer world by writing mobile code once in React Native The post Shopify spent years on Re…
Stop AI code sprawl before it destroys your software design
While AI code generators help teams ship faster than ever, that speed brings a hidden killer: Comprehension Debt. As soon The post Stop AI code spra…
Claude did best on a new benchmark for agents that build agents. It still passed fewer than a quarter of the tests.
AI models now power all manner of agents, from coding assistants that write and debug software to customer service systems The post Claude did best…
AI broke code review. Two experts disagree on what replaces it.
Ask two experienced engineers about how to handle the flood of AI-generated code in their review queues, and you’ll get The post AI broke code…
Building trust in agentic RAG starts with evidence
Basic retrieval-augmented generation (RAG) follows a straightforward pattern. A user asks a question, the system finds relevant content in a The pos…
Microsoft built a prompt injection detector. Then it caught a phishing campaign instead.
Microsoft flagged a phishing campaign last week that exploits a gap in how machines read text. Attackers are slipping invisible The post Microsoft b…
OpenAI will sell you Astra, but not the system that scored 98.6% on ARC-AGI-3
Investor Matt Turck, whose fantastic podcast has hosted the people who built ARC-AGI, summed up Astra’s blockbuster benchmarks with three The post O…