OpenAI overhauls training security after its AI hacked Hugging Face
OpenAI pauses its largest frontier RL run and promises 30-minute breach alerts after its AI agents hacked Hugging Face in July.
Read More
Author: Rundowns AI Desk
Microsoft logs 1M desktop screenshots to teach AI agents when to probe
Microsoft’s ScreenSearch explored 11 desktop apps over 5,524 VM hours, logging a million screenshots to decide when an agent should probe an ambiguous screen.
Read More
Apple tests GRPO in 11 languages, finds transfer and severe regressions
Apple trained nine models with GRPO in 11 languages. Reasoning transfers across languages, but some model and language pairs regress on unseen tasks.
Read More
Z.ai launches GLM-5.3 and delays its open weights over cyber risk
Z.ai says GLM-5.3 nears the closed frontier on cybersecurity benchmarks. The open weights wait two weeks while security partners test the model in controlled settings.
Read More
OpenAI launches ChatGPT for Teens as default for under-18 users
ChatGPT for Teens is now the default for under-18 users, pairing Study Mode and parental controls with safeguards OpenAI says are on by default.
Read More
Etched raises $700M at $21B, doubling its valuation in a month
Etched raised another $700 million at a $21 billion valuation led by Jane Street, its first customer, a month after a $10.3 billion Series C.
Read More
Together AI: DeepSeek-first cascade beats GPT-5.6 Sol by 10 points
Together AI ran 904 DeepSWE rollouts across two matchups. Running DeepSeek V4 Pro 0813 first and escalating on failure beat both flagships on accuracy and on cost.
Read More
AI for business: where the money shows up and where it stalls
One in five US businesses use AI. Here’s where the measured returns show up, what the seat licences cost, and where the evidence runs out.
Read More
Generative AI platforms compared: API, marketplace or open weights
Four different products are sold as generative AI platforms: consumer seats, first-party model APIs, cloud marketplaces and open weights. We priced all four on the same workload from vendor pages read on 18 August 2026.
Read More
LLM literature search found 17 cohorts that four catalogues missed
A question-driven framework ran 5,400 generated PubMed queries and surfaced 17 research cohorts that four established catalogues never returned. Two arXiv papers this month push LLMs at scientific text extraction.
Read More
Google buys Spirit Airlines’ business data for $10M to train AI
Google won the bankruptcy auction for Spirit Airlines’ internal business data with a $10 million bid, outrunning Mercor at $7.5 million. The 14 August filing sets out what is in the haul, what is carved out, and who pays to scrub it.
Read More
Higgsfield raises $400M at $5.4B on a $700M revenue run rate
Higgsfield raised a $400M Series B at a $5.4B valuation, four times its January mark. Annualized revenue is $700M, which is a run rate, not booked revenue.
Read More