World models represent a new AI category, learning world dynamics from visual and auditory experience.
This enables advances in robotics, gaming, advertising, defense, education, and retail.
Here's @jeffrey_hawke of @odysseyml at @raais
Show more
Breakthroughs emerge not from scaling or data alone, but from incorporating human feedback on preferences, a key lesson from LLMs like GPT-2 applied to world models. This preference learning is an underexplored frontier.
Here's Fabian Gura of @odysseyml at @raais 2026
Show more
always important to think for yourself
the models have no moat (OpenAI, Anthropic, XAI)
the IDEs have no moat (Cursor, Windsurf)
the harnesses have no moat (Cognition, Factory, LangChain)
the app builders have no moat (Replit, Lovable, Bolt)
the wrappers have no moat (Harvey, Abridge, OpenEvidence)
the inference providers have no moat (Together, Fireworks, Groq)
the voice layer has no moat (Sierra, Decagon, ElevenLabs)
the data labeling companies have no moat (Scale, Surge, Mercor)
the AI infrastructure has no moat (Baseten, Modal, Railway)
the neoclouds have no moat (CoreWeave, Lambda, Crusoe)
the generative media companies have no moat (Runway, Higgsfield, Suno)
apparently nobody in AI has a moat
except the venture firm ☠️
Show more
china bout to accelerate more
every lil bit helps
World models, trained to predict images and videos, will unlock understanding of the physical world beyond text.
Expect major advances in robotics, gaming, and unforeseen applications according to @JonathanSadeghi of @odysseyml at @raais :)
Show more
see what i mean? serious regression in performance
is it just me or is opus 5 (on high) making a lot of mistakes when you ask it are you sure or point out something that seems off
i'm getting a lot of this in the last weeks, ro eg
- "Confirmed, and I need to correct myself"
- "Let me fix the note, which carries the same overreach."
- "No — and I was wrong to imply otherwise. Thanks for stopping on it."
Show more
great for margins
Eventually 99% of agentic tokens will just be the cached prompt.
by “next” you mean “current” right?
is it just me or is opus 5 (on high) making a lot of mistakes when you ask it are you sure or point out something that seems off
i'm getting a lot of this in the last weeks, ro eg
- "Confirmed, and I need to correct myself"
- "Let me fix the note, which carries the same overreach."
- "No — and I was wrong to imply otherwise. Thanks for stopping on it."
Show more
does your iphone store duplicates of whatsapp photos/videos in google photos/apple photos too these days?
super annoying
i didn’t have two rounds of broken shoe laces at critical moments in the match on my bingo card
and this is adidas no less
seems a little suss eh
and maybe that’s ok? 🙊
I keep coming back to this 42% number.
Of the 892 unicorns founded by European-origin founders, 371 were built in the US.
European talent remains one of America’s biggest startup advantages.
Show more
mood in europe right now when the chinese uber eats ships better models than you do and your best hope prefers to serve them rather than cook
literally cooking @angles_hq
Day 1 (of many) of getting @angles_hq into the App Store
Sneak peek of some of our BTS / work with @afterimage
Will be sharing more soon on how we’re designing final polish + onboarding to get angles out to everyone :-)
Show more
total air defense is a non-negotiable in today’s world
i get it’s super expensive and super complicated to build but surely more neoprimes should work on this
pray for a @thsottiaux reset
phones blaring mental in the uk
wildfires, not nukes.
kinda wild how private equity-backed companies spend 90% less on AI per employee than venture-backed companies
when pe-backed firms are the ones that'll get eaten
the fact that founders of ai labs so regularly leave the companies they started, often within months is raising or not long thereafter before anything has really happened, should somehow register as a reward signal that:
a) you should think really hard about your motivations to start a company
b) doing so should be because there’s just no other vehicle for you to express your talent and conviction on bringing something new into the world and doing whatever it takes to win
c) you have to sure about your choice of cofounders, come rain come shine
otherwise, just don’t do it, and that’s fine!
Show more
world models are a continuously adapting training environment for ai
here’s @olivercameron from @odysseyml
