News, p. 20
News
427 stories · page 20 of 48JetBrains' Junie now runs entirely on your Mac, no cloud, no tokens
JetBrains' code agent now runs locally. No credits, no quota, the code never leaves the machine. The price is the hardware: an M5 chip and at least 64 gigabytes of memory.
Read →JetBrains picked the older Qwen 3.6 over 3.8 because the newer model is slower on a laptop
Behind local Junie sits a string of engineering trade-offs: a quantized model, reasoning turned off, remembered context. The most interesting one is the model choice. The newer Qwen 3.8 was dropped because, with reasoning on, it churns out around five times more tokens.
Read →GPT-5.6 lands in AWS's Kiro, and the headline number is cost
The whole GPT-5.6 family, that is Sol, Terra and Luna, now works inside Kiro, AWS's spec-driven coding agent. OpenAI and AWS tuned the environment together. The number they put front and center isn't about quality. It's about cost.
Read →Vercel Sandbox goes from one region to four, and Paris is the first in Europe
Vercel Sandbox has gone global, and for a start it covers four cities, one of them Paris, the first European address on the list. Region selection is available on every plan, but only Pro and Enterprise get a fallback region on failure.
Read →ElevenLabs shipped a command line written for agents, not people
ElevenLabs' new command line puts their whole product into the terminal: music, text to speech, agents. Responses come back as structured JSON, and errors are written so a program can parse them. Its user is another agent. The human comes second.
Read →Mistral and HUMAIN are building sovereign AI for Saudi Arabia
Mistral and Saudi Arabia's HUMAIN announced joint work on sovereign artificial intelligence for Saudi Arabia and the Middle East: infrastructure, models, deployment, in the range of hundreds of millions of euros. The computing stays inside HUMAIN's data centres.
Read →DeepSeek shipped vision over the API and counts an image as up to 384 tokens
An experimental variant of their fast model now understands images too, not only text. You pay for it at the text model's rate. For the promised jump on the multimodal agentic tests there is not a single published number.
Read →Microsoft corrected its own field: the 10.0 hole in Entra ID was not exploited
On 20 August Microsoft announced a critical hole in Entra ID with the maximum score of 10.0 and noted in its record that it was being exploited. A day later it corrected the field. The patch is entirely on their side, and for customers there is nothing to do.
Read →AWS Glue 6.0 cuts the price by 30 percent and adds full Iceberg v3 support
AWS shipped version 6.0 of Glue, its data-processing service, at 30 percent lower cost than previous versions. Underneath is a new engine, and on top, full support for the Iceberg table format v3.
Read →