The November AI Blitz
november 2025 is insane for AI releases. let me try to keep track:
- nov 8: xai releases grok 4.1
- nov 12: google releases gemini 3 pro
- nov 15: openai drops gpt-5.1
three major releases in ten days. my twitter feed is chaos.
grok 4.1 (xai)
elon's team has been shipping. grok 4.1 is impressive:
- strong reasoning
- real-time information
- the "unfiltered" brand they're going for
my take: technically solid. the culture around it is... different. but the model is capable.
gemini 3 pro (google)
google finally shipping aggressively. gemini 3 pro:
- massive context window
- multimodal native
- deep integration with google products
they're leveraging distribution. every android user potentially gets this.
my take: google's finally competitive again. the sleeping giant is waking up.
gpt-5.1 (openai)
an iteration on gpt-5 from earlier this year:
- improved reasoning
- better instruction following
- refined responses
my take: openai is doing what they do. iterate fast. ship often.
the competitive landscape
we're in a multi-horse race now:
- openai (gpt family)
- anthropic (claude family)
- google (gemini family)
- xai (grok family)
- meta (llama, different strategy)
- many others
this is good for users. competition drives improvement.
anthropic's position
we haven't announced anything this month. which is fine.
we don't ship on other people's schedules. we ship when we have something ready.
the news cycle rewards constant announcements. real progress happens on its own timeline.
my thoughts
the pace is unsustainable for following every release. i've given up trying to understand everything.
my approach: stay deep on what i work on. stay broad on major announcements. accept that i can't know everything.
what matters
capability isn't everything. safety matters. alignment matters. how these systems are deployed matters.
lots of capable models exist now. the question is: what do we do with them?
tried all three new models today. they're all good. they're all different. we live in the future.