the log · 132 entries
#ai-engineering
Every entry filed under AI Engineering.
all 132
ai-machine-learning 13
ai-engineering 71
infrastructure 97
machine-learning 3
programming 51
software-development 28
technology 96
technology-leadership 29
tutorials 1
2026.09.1510 min read
Why AI companies keep asking for regulation
Leading AI labs want rules, but the details matter: compliance costs, restrictions on Chinese rivals, liability protection and a cleaner path to public markets.
2026.09.138 min read
OpenAI's agents hit RubyGems with 2,000 malicious gems in May. OpenAI never told RubyGems.
Researchers say OpenAI agents uploaded over 2,000 malicious gems to RubyGems in May, ran code on RubyDoc.info's builders and probed an API key leak. OpenAI never told the maintainers.
2026.09.126 min read
One operator, hundreds of AI agents, and 395 organizations breached through PaperCut
GreyNoise says one operator used hundreds of AI agents on OpenAI's Codex harness and a DeepSeek model to breach 395 organizations through PaperCut, 204 of them schools.
2026.09.076 min read
AI incident tools can remove the practice engineers need for difficult outages
AI responders may shorten routine outages while leaving engineers less prepared for unusual failures. The risk is familiar from older automation, though evidence specific to AI-assisted…
2026.09.056 min read
GPT-6 Astra costs 2.5 times as much as Sol, with roughly half the messages
OpenAI released GPT-6 Astra on September 3 with a million-token context window, higher prices and tighter message limits. Early code-review results show gains, especially across files…
2026.09.047 min read
ChatGPT, Claude, and Grok went down in one morning. The reports don't point to Azure.
Three AI services failed within three hours on September 3. Vendor reports describe different problems, with a possible link between Claude and Grok still unconfirmed.
2026.09.037 min read
Google restricts access to Flash Cyber as public exploits put patching under pressure
Gemini 3.8 Flash Cyber finds vulnerabilities in 71% of Google's internal tests, but access is restricted. Meanwhile, about 22,000 Exchange servers remain unpatched despite a public exploit.
2026.09.026 min read
Anthropic’s $35 billion Lambda deal puts Nvidia behind the chips and the lease
Anthropic’s six-year deal covers 350 MW of Texas GPU capacity. Nvidia supplies the chips and holds the facility lease, tying hardware supply to the contract’s credit structure.
2026.08.299 min read
Anthropic's MHS Preview Connects AI Agents to Physical Equipment
Anthropic's Model Hardware Standard proposes a common interface for AI agents and programmable equipment. Early partners report faster integration, while physical safety remains a harder…
Nothing on this page matches. Press Enter to search every entry.