1 November 2026
Anthropic's Cursor extension was backdoored as AI funding worries grow
Security stories led the day. A backdoored Anthropic extension, malware that uses LLMs and a ChatGPT account takeover write-up all touch tools many teams use. Funding news for OpenAI and Anthropic also raised questions about the cost of AI.
Security and safety
Anthropic VS Code extension on Cursor backdoored for nearly 12 hours (opens the source in a new tab)
Ars Technica reports that Anthropic's extension on Cursor was backdoored for nearly 12 hours. It is still unknown how attackers got into Anthropic's verified Cursor account.
Why it matters: If your team uses Cursor, check which extensions are installed and ask your developers whether anyone updated this one.
Ars Technica · 30 Oct
How researchers took over ChatGPT accounts (opens the source in a new tab)
A write-up on parameter.ai describes how its authors took over ChatGPT accounts.
Why it matters: Use a strong password and two-step login on every ChatGPT account your team uses, and watch for any fix from OpenAI.
parameter.ai via Hacker News · 31 Oct
Malware uses LLMs to rank victims by how much they're worth (opens the source in a new tab)
Cisco Talos reports on malware that uses large language models to rank victims by their worth.
Why it matters: Attackers may pick targets more carefully, so keep software updated and train staff to spot odd emails and links.
blog.talosintelligence.com via Hacker News · 31 Oct
AI startup reportedly hacks Anthropic and reaches employee details (opens the source in a new tab)
India Today reports that an AI startup hacked Anthropic and gained access to employee details and internal messages.
Why it matters: Treat this as an early report, and avoid sharing sensitive client data with any AI vendor that you have not checked.
indiatoday.in via Hacker News · 31 Oct
Tools and products
ChatGPT and Codex usage limits are merging (opens the source in a new tab)
An OpenAI help page says ChatGPT and Codex usage limits are merging.
Why it matters: If you use both, check your plan limits, because heavy use of one may now affect the other.
help.openai.com via Hacker News · 31 Oct
Claude now blocks video game cloning and reverse-engineering (opens the source in a new tab)
Tom's Hardware reports that Anthropic's Claude now blocks requests to clone video games and reverse-engineer them.
Why it matters: If you build software with Claude, expect it to refuse some copying and reverse-engineering requests.
tomshardware.com via Hacker News · 31 Oct
Post claims Anthropic blocks users over training larger models (opens the source in a new tab)
A post on Twitter says Anthropic blocks users for anything touching training larger models. This comes from the post alone.
If you use Claude for model training or data work, read Anthropic's usage rules so your account is not blocked.
Enterprise
OpenAI raises less than half of its $30B target (opens the source in a new tab)
The Wall Street Journal reports that OpenAI raised less than half of its $30B funding target.
Why it matters: If you build on OpenAI, keep a backup option in case prices or plans change.
wsj.com via Hacker News · 31 Oct
Anthropic may get a junk credit rating after IPO (opens the source in a new tab)
The Financial Times reports that Anthropic is expected to get a high junk rating after its IPO. TechCrunch notes this is only Blackstone's forecast.
Why it matters: Costly borrowing could squeeze AI vendors, so avoid locking your business into one provider.
TechCrunch · 30 Oct
Nvidia weighs cosigning Nscale's West Virginia AI data center (opens the source in a new tab)
TechCrunch reports that Nvidia is considering cosigning a loan for Nscale's AI data center in West Virginia.
Why it matters: It shows how much AI capacity depends on borrowed money, which can affect the prices you pay later.
Research
Study looks at generative AI and critical thinking (opens the source in a new tab)
A 2025 paper in the ACM library looks at how generative AI affects critical thinking, using self-reported answers.
Why it matters: Ask your team to review AI output before it goes to clients, and note that this study relies on self-reports.
dl.acm.org via Hacker News · 31 Oct
GPT-5.6 Sol can beat Factorio on peaceful mode (opens the source in a new tab)
A blog post says GPT-5.6 Sol can beat the game Factorio on peaceful mode.
Why it matters: It hints that AI can plan long, multi-step tasks, which matters if you want agents to run longer jobs.
ryanmadden.net via Hacker News · 31 Oct
AI text watermarking and quality loss (opens the source in a new tab)
A blog post looks at AI text watermarking and how it can lower text quality.
Why it matters: If AI text is marked in future, check that your copy still reads well.
blog.keyvan.net via Hacker News · 31 Oct
Also worth knowing
Judge lets NAACP lawsuit over xAI pollution go ahead (opens the source in a new tab)
A judge allowed an NAACP lawsuit to move forward. It accuses the SpaceX-owned AI company xAI of breaking the Clean Air Act over its data center.
Why it matters: Legal pressure on data centers could add cost for AI providers, so keep an eye on your vendors' pricing.
TechCrunch · 31 Oct
Computer science programs lose students as AI replaces jobs (opens the source in a new tab)
Bloomberg reports that computer science programs are losing students as AI replaces jobs.
Why it matters: Technical hiring may get harder, so plan how you will find or train people who can run your AI tools.
bloomberg.com via Hacker News · 31 Oct