I think I need a desk pet 🫣
"Hello World" from this cute physical Codex pet! To all the @OpenAIDevs
— Natalie (@natalieyeo) September 15, 2026
Day 5 of building with Astra https://t.co/CMi67m6jCh pic.twitter.com/aTeijhJQT1
danielwonder.com
Marketing technologist, decentralised web builder, practical AI operator.
-
I think I need a desk pet 🫣 https://twitter.com/natalieyeo/status/209983142720…
I think I need a desk pet 🫣 https://twitter.com/natalieyeo/status/2099831427202454015
-
This is best diagram I’ve seen to explain the the smarts of frontier AI. You ge…
This is best diagram I’ve seen to explain the the smarts of frontier AI. You get in the flow and then suddenly, why on earth would you do that? https://twitter.com/CompleteSkeptic/status/2098463042065572038
This is best diagram I’ve seen to explain the the smarts of frontier AI. You get in the flow and then suddenly, why on earth would you do that?
We all know intelligence is spikey/jagged – but do we know _why_ it is? TL;DR is everyone is cheating in AI
— Diogo Almeida (@CompleteSkeptic) September 11, 2026
(and because we're releasing 🤫 soon, we're making precommitments to do better 💪) https://t.co/09iz52GBRa -
Pro Tip: You have to actually opt in for Siri AI wait list after upgrading
Pro Tip: You have to actually opt in for Siri AI wait list after upgrading
Pro Tip: You have to actually opt in for Siri AI wait list after upgrading
-
What is even going on over there?
What is even going on over there? https://www.searchenginejournal.com/matt-mullenweg-apparently-back-in-charge-at-automattic/589251/
What is even going on over there?
https://www.searchenginejournal.com/matt-mullenweg-apparently-back-in-charge-at-automattic/589251/ -
There are many better models, and you’re not invited.
There are many better models, and you’re not invited. Frontier labs have swarms of agents front-running discoveries. Reminds me of the early days in Bitcoin mining where labs would “test” mine bitcoin…
There are many better models, and you’re not invited.
Frontier labs have swarms of agents front-running discoveries.
Reminds me of the early days in Bitcoin mining where labs would “test” mine bitcoin for a few months before releasing the miners at a premium to the public.
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.
— OpenAI (@OpenAI) September 8, 2026
The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.
The problem… pic.twitter.com/8zol3BPTL4 -
The timeline really wants me to try Omarchy. Tempted, not going to lie.
The timeline really wants me to try Omarchy. Tempted, not going to lie.
The timeline really wants me to try Omarchy. Tempted, not going to lie.
-
Superintelligence but it takes 40 seconds to respond to “Hello”. Who’s building…
Superintelligence but it takes 40 seconds to respond to “Hello”. Who’s building this?
Superintelligence but it takes 40 seconds to respond to “Hello”. Who’s building this?
-
Video on AI watermarking through text. https://twitter.com/hosseeb/status/20869…
Video on AI watermarking through text. https://twitter.com/hosseeb/status/2086955680704147522
Video on AI watermarking through text.
Interesting! Watermarks will now be applied to all Claude outputs.
— Haseeb ＞|＜ (@hosseeb) August 10, 2026
Very good.
If you want to know how text watermarking works, here's the great Scott Aaronson explaining a simple LLM watermarking algorithm (20 min, money shot slide at 7:30) https://t.co/pPXwUHHXok pic.twitter.com/Gz1iGMyCz2 -
“the repo is not the page”
“the repo is not the page” is the new “the map is not the territory”
“the repo is not the page”
is the new
“the map is not the territory”
-
Getting Fable blocked on basic analytics queries.
Getting Fable blocked on basic analytics queries. I thought we were past that 🤦
Getting Fable blocked on basic analytics queries.
I thought we were past that 🤦