My guess is that the fraction of humans who understand enough about LLMs to ask those questions overlaps pretty heavily with the fraction of humans who know that LLMs are almost universally terrible at accurately answering questions about themselves
RESEARCH
-

Scott McCloud illustration referenced in world models research paper
By
–
Haha that was the same Scott McCloud illustration I also referenced in a paper about world models https://
arxiv.org/pdf/1803.10122 https://
worldmodels.github.io -

GPT-5 Parameters Revealed: 3-5T According Samsung Executives
By
–
Who remembers this from September 2024? GPT-5: either 3-5T parameters or 3.5T, according to Samsung execs
-
Automated reasoning switches and artifact pollution in AI systems
By
–
Yes, the most naive way of doing this is an automated reasoning switch like old Claude or KAT-V1 (
https://
arxiv.org/abs/2507.08297). But artifacts from reasoning always pollute non-reasoning tasks. This is a bigger issue imo. -
Google DeepMind Unveils Genie 3 Real-Time 3D World Generation
By
–
Hier Google a dévoilé l'une des prouesses technologiques les plus hallucinantes jusqu'ici.
— VISION IA (@vision_ia) 6 août 2025
Ceci à été créé en temps réel avec le Genie 3, la nouvelle IA de Google Deep Mind.
Il est même contrôlable ! Un monde virtuel 3D, au réalisme fou.
La matrice est en train de craquer pic.twitter.com/q6zzJCGiGiToday Google unveiled one of the most mind-blowing technological feats to date. This was created in real time with Genie 3, the new AI from Google DeepMind. It's even controllable! A 3D virtual world, with insane realism. The Matrix is starting to crack.
-
Living at a Pivotal Moment in Technological History
By
–
Amazing that we happen to be alive at this moment in time
-
SciSpace Agent AI Co-Scientist Now Available on Product Hunt
By
–
.@SciSpace_ Agent, the #AI Co-Scientist built for researchers, is now live on Product Hunt: https://t.co/sO4j5uclnb
— Kirk Borne (@KirkDBorne) 6 août 2025
…SciSpace Agent unites hundreds of academic tools and databases in one platform. With a single prompt, it reads literature, analyzes data, drafts papers,… https://t.co/WTwzMTgI1P.
@scispace Agent, the #AI Co-Scientist built for researchers, is now live on Product Hunt: https://
producthunt.com/products/types
et/launches/scispace-agent
… …SciSpace Agent unites hundreds of academic tools and databases in one platform. With a single prompt, it reads literature, analyzes data, drafts papers, -

Kaggle Game Arena AI Chess Tournament with Magnus Carlsen Commentary
By
–
The #Kaggle Game Arena's inaugural Live #AI Chess Exhibition Tournament features models from top AI labs, and commentary from world-renowned chess players, plus a live-stream of the championship match-up and tournament recap from @MagnusCarlsen at: https://
youtube.com/@TakeTakeTakeA
pp
… -
Kaggle Game Arena: AI Models Compete in Strategic Games Benchmark
By
–
Check out the new @Kaggle Game Arena public benchmarking platform where top #AI models compete head-to-head in strategic games to provide a more robust and dynamic measure of their true capabilities: https://
kaggle.com/blog/introduci
ng-game-arena
… Learn more here: -
Updating Mental Models of Large Language Models
By
–
This is a big thing that is (mostly? increasingly?) true and which people haven't updated their own mental models of LLMs to take into account.