Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-09-04 · high
Downdetector logged more than 35,000 US reports for ChatGPT alone, and no vendor has confirmed a shared cause.
Why now: The outages happened September 3 and recovery came by early afternoon, with the shared-cause question still open.
ChatGPT, Codex, Claude and Grok all reported outages starting around 7:57 AM PT on September 3. The Register's report is corroborated by Axios and the vendors' own status pages. Downdetector logged more than 35,000 US reports for ChatGPT, 1,400 for Claude and 1,200 for Grok, and all four recovered by 12:38 PM PT.
Anyone running a product where an LLM call sits in the request path watched every fallback option go dark inside the same two-hour window. Losing one vendor for an afternoon is a routing problem. Losing four at once is a different kind of failure.
Nobody has confirmed why. Cloudflare says it saw no significant disruption on its end, ruling out the obvious shared-infrastructure explanation. xAI traced its own outage to a Memphis data center, an internal cause specific to Grok rather than something upstream that would also hit ChatGPT, Claude and Codex. The Register's reporting doesn't establish a common root cause, so this could be four unrelated failures that happened to overlap by the clock.
Multi-vendor fallback assumes each provider fails independently. September 3 is a case where that assumption didn't hold, whatever the eventual explanation turns out to be.
Each link below shares sources, entities, or timing with this story.
Starting around 7:57 AM PT on September 3, all four reported outages simultaneously, with Downdetector logging 35,000+ US reports for ChatGPT, 1,400 for Claude and 1,200 for Grok before recovery by 12:38 PM PT. Cloudflare denied any significant disruption and xAI traced its ow...
Two of the 11 bugs paid real bounties, $10,000 from Microsoft and $3,133.70 from Google.
Code freezes July 29, the repo archives August 10, and Flowise hasn't said how cloud customers export their data.
67 on coding against Fable 5.1's 70 in Claude Code. Astra does post a 2% hallucination rate against 9.4% for GPT-5.6 Sol, and 0% scope violations against 48%. Per-task cost runs the other way, $4.72 for Astra against $9.18 for Fable 5.1 at identical $10/$50 list pricing, and A...
The general resolution ballot ranges from Proposal A's total prohibition via Social Contract amendment, needing a 3:1 majority, and Ian Jackson's Proposal C ban, through Proposal B's six-condition permission regime, to Proposal H arguing LLM usage accelerates ecological destru...
Court filings unsealed this week show Google won a bankruptcy auction for Spirit Airlines' internal corpus. The Register has the itemized manifest, corroborated by Axios, CNN Business, Bloomberg Law, and Skift: Roughly 100 million emails. 500 million Microsoft Teams items. 17...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.