We’ve been measuring AI wrong; why economically valuable work is the new benchmark
As the AI industry gradually builds standardization guidelines and systems, such as those overseen by the Tokenonmics Foundation, the need The post We…
Programming
⚑ Report a ProblemLatest Programming news from Tech News
All topics
AI
agents
ai
api
architecture
automation
aws
beginners
career
claude
database
devchallenge
devops
javascript
learning
llm
machinelearning
mcp
opensource
performance
productivity
programming
python
react
rust
security
showdev
tutorial
typescript
webdev
Fable 5 vs Opus 4.8: The real stakes, not the spec sheet
This week Anthropic finally dropped the first model in its Mythos-class tier, Fable 5. The pitch was clear: this is The post Fable 5 vs Opus 4.8: The …
Cleaner AI training data, fewer bugs: Sonar’s SonarSweep explained
Large language models have moved quickly from novelty to daily infrastructure in software development. We are no longer using AI The post Cleaner AI t…
Google’s DiffusionGemma is 4x faster than its other Gemma models
About a year ago, Google demoed a diffusion model at its I/O developer conference, but went quiet about the technology The post Google’s Diffusi…
Fable 5: Guardrails and burn rate are annoying users, who say it’s still better than Opus 4.8
On Tuesday, Anthropic debuted Fable 5, the first — and much-anticipated — generally available Mythos-class model.  Anthropic say…
Why Anthropic just doubled Claude Cowork limits at no charge
Anthropic has announced a limited-time promotion that doubles users’ five-hour usage limits in Claude Cowork. Announced on Monday, it’s su…
Google Gemma 4 12B nearly matches 26B benchmarks — and runs on your laptop
Google has introduced Gemma 4 12B, a new model designed to bring high-performance, multi-modal intelligence to standard laptops. Small enough The post…
OpenAI’s Codex adds new tools — Sites, Annotations, more plugins — for knowledge workers
As OpenAI announced on Tuesday, 20% of Codex’s 5 million weekly active users are now knowledge workers and not coders. The post OpenAI’s C…
Why GPT-5.4, Claude, and Gemini can’t agree on basic, real-world facts
As the frontier model race accelerates, AI devotees are splitting their loyalty across the major providers at both the user The post Why GPT-5.4, Clau…
“Tokenmaxxing is real, expensive & it’s spreading”: New tools emerge to stop AI budgets from exploding
There’s a new weapon in the fight against tokenmaxxing. Tokenmaxxing, of course, occurs when an enterprise decides that AI token The post “…
Google ranks the best AI for building Android apps, and the winner isn’t Gemini
Google wants software developers to use the best possible AI models when building Android applications; consequently, the company debuted its The post…
Anthropic hires OpenAI co-founder Andrej Karpathy to lead Claude pre-training research
Anthropic scored a major hire today. Former Tesla senior director and OpenAI founding member Andrej Karpathy is joining the organization The post Anth…
Why production RAG systems give confident, wrong answers at scale
In production RAG systems, the biggest bottleneck usually isn’t the LLM. It’s retrieval. Most teams start with a simple pattern: The post …
I tested OpenAI’s three claims about GPT-5.5 Instant, and only one fully held up
Last week, OpenAI replaced GPT-5.3 Instant as ChatGPT’s default model with GPT-5.5 Instant, rolling it out to all users for The post I tested Op…
Anthropic trains Claude to resist blackmail & self-preservation behavior via agentic misalignment
Anthropic doubled down on the fight against agentic misalignment on Friday, the mechanics of which could cause AI models to The post Anthropic trains …
Anthropic will let its managed agents dream
In April, Anthropic launched the public beta of Managed Agents, its platform for running AI agents on its infrastructure. On Wednesday, The post …
The company that made RAG mainstream is now betting against it.
Pinecone just declared the RAG era over. Pinecone built the vector database category. It defined RAG as the standard pattern The post The company that…
How NetEase Games cut LLM cold starts from 42 minutes to 30 seconds
At NetEase Games, we learned a hard lesson about large language model (LLM) inference in production: elastic compute is only The post How NetEase Game…
Why JSON Schema matters more than ever in the age of generative AI
As enterprises grapple with the unpredictability of large language models, the quietly ubiquitous JSON Schema standard is emerging as a The post Why J…