All I see is regret-minutes everywhere
Elon Musk@elonmusk13 Dec 2022United States!
United StatesConnected via United States App StoreLocation may be affected by a proxy or VPN.Account-level information, not a live location or per-post device.
Replying to @EMTSLA @HarryStebbings
Unregretted user-minutes is the metric that matters most
Replying to @GergelyOrosz
This is basically what compilers did to programmers in the 70s. Assembly didn’t go away but the overall complexity increased and the job moved up a level
cd retweeted
GenAI kicked off the golden age of the Socratic Method. Millions now ask, learn, improve the question, repeat
cd retweeted
Replying to @naval
Van Gogh is the ultimate example of great product and zero distribution
cd retweeted
xAI is hosting a Grokathon and this time in London on Jan 17, 2026
A 12‑hour, non‑stop coding sprint to build real applications with Grok and show what’s possible
And this time, there will be 3 winners
Tell your friends to join...
cd retweeted
I would like some help with testing an Automated distrubution AB Test. If you can spare a moment to click through to help me determine if my AB Test(s) are working, that would be incredible!
d2s.io/7KiAIz
Thanks! :-)
#buildinpublic
cd retweeted
GenAI kicked off the golden age of the Socratic Method. Millions now ask, learn, improve the question, repeat
cd retweeted
Replying to @tszzl
Nvidia doesn’t need to believe in AGI. They need to sell to those who believe
cd retweeted
Replying to @gabriel1
Motivation comes and goes. Structure is what keeps things moving. Like Picasso said “inspiration exists but it has to find you working”
cd retweeted
Replying to @SawyerMerritt
Yes it’s a bubble. But that’s how new industries get built. Many AI startups will fail, but the ones that survive will be giants.
cd retweeted
Replying to @yudapearl
Distinguishing true thinking from the illusion of thinking is not that important when you are just using ChatGPT. But it becomes absolutely critical when we start designing AI-native systems that assume intelligence where there is none.
cd retweeted
Replying to @r0ck3t23
LLMs are trained on pattern completion, not correctness. They don’t learn facts, they learn what usually comes next.
cd retweeted
Replying to @haider1
LLMs are a “momentary fixation” the same way humanoid robots are. We are fixated on human-like forms, but evolution didn’t choose the human form for efficiency, just survival.





