BIZN.org
BIZN.org
Four items that matter to a small business, with sources. News moves fast and vendor announcements are vendor claims — check the linked source before acting on any of it.
Apple announced on its China newsroom on October 31 that Apple Intelligence is coming to iPhones in mainland China — with Alibaba's Qwen model, rather than ChatGPT, powering some requests. Testing starts in the iOS developer beta on November 4, with the public release in iOS 27.3 in December. Qwen plugs into Siri for certain requests, plus Writing Tools and Visual Intelligence, so users can ask about photos, PDFs and presentations and generate original content. The first wave is iPhone-only — iPhone 15 Pro and Pro Max, iPhone 16 and later, iPhone Air and iPhone Duo — with device and Siri language set to English or Simplified Chinese, and it does not include Siri AI or the other next-generation features announced at WWDC 2026. Alibaba says participation is opt-in and that requests and responses will not be used for training.
Why it matters: If you sell to, or make content for, customers in China, Apple's AI now answers with a different model depending on the region — Qwen's answers will not match ChatGPT's, so test in the November 4 beta rather than assuming your prompts behave the same. The opt-in, no-training commitment is also the privacy line to hold Apple and Alibaba to if your staff or customers use these features for business documents.
9to5Mac: Apple confirms iOS 27.3 beta is coming next week with Apple Intelligence in China →
S&P Global Ratings moved its outlook on Meta to negative from stable on October 30, keeping the AA- rating, Reuters reported — citing AI spending and negative free cash flow, with S&P forecasting free operating cash flow deficits of $22 billion in 2026, $71 billion in 2027 and $39 billion in 2028. Meta pushed back, saying it disagrees with S&P's view of its AI monetization. Separately, the Financial Times reported that Blackstone — anchor investor in a $63 billion deal financing Anthropic's chips — now expects Anthropic to be rated junk, or speculative grade (BB), after its IPO, having previously expected an investment-grade rating. That is an investor's forecast, not an agency rating: no agency has rated Anthropic yet, and its IPO is expected in mid-November at a reported target valuation of around $2 trillion.
Why it matters: A negative outlook is a warning, not a downgrade — Meta's AA- is still one of the strongest corporate ratings. But the pattern is the story: in the same week OpenAI reportedly cut its funding target, lenders are repricing AI risk across the board. For a small business, vendor stability is now a fair procurement question: ask what pricing is contractually committed and for how long, rather than assuming today's subsidized prices hold.
Seoul Economic Daily: Wall Street Turns Cold on AI: Meta Outlook Cut, Anthropic Seen as Junk →
Blockchain intelligence firm Chainalysis disclosed on October 31 that an external attacker used what it calls "offensive agentic techniques" to break into its internal systems earlier in October. The data accessed was customer user IDs, full names, email addresses and organizational role labels. The company says its investigation has found no evidence of unauthorized access to investigation content, authentication credentials or session tokens, and says the incident is unrelated to the recent $6.3 billion Binance theft. It has not said which vulnerabilities were exploited, how long the attacker was inside, or how many users were affected; customers were told to enable two-factor authentication and watch for unexpected password-reset requests and scammers impersonating Chainalysis.
Why it matters: This is the second AI-security story in a week, after the tampered Cursor extension — and the lesson is the same: "non-sensitive" data is exactly what precision phishing runs on. Names, work emails and job titles, correctly matched, make a scam message far more convincing. If your staff's names and roles are public, use this as the prompt to confirm two-factor authentication is on everywhere and to warn the team that the next phishing attempt may address them by name and title.
DeepSeek's free window for Search and Fetch tool calls ends today. From 00:00 Beijing time on November 1 (16:00 UTC on October 31), successful calls to its Search and Fetch APIs cost $0.75 per 1,000 requests, according to DeepSeek's pricing page. The change reaches coding-agent users too: server-side tool calls made through DeepSeek Harness, Claude Code or Codex on a DeepSeek key were free until the cutoff and now bill at the same rate, on top of token charges. Failed, empty and timed-out calls are not charged.
Why it matters: If your team runs coding agents on DeepSeek keys, a search-heavy agent has quietly gained a new line item — 2,000 searches a day adds $1.50 a day before any tokens. Check your agent's tool-call logs and set a spend cap; the charge comes out of the same prepaid balance as token usage.
Limit Digest: DeepSeek Search API pricing: free tool calls end Oct 31 →
For what the AI companies and their leaders are saying on social media — claims and opinions, not verified news — see The Social Wire →
Browse the full directory → Back home →
This day has a permanent page: November 1, 2026. Every day is kept — see the News archive. The News page always shows the latest day.
← Latest day · News archive · Previous day: October 31, 2026