Scaling is over, the bubble may be deflating, LLMs still can’t reason, and you can’t trust Sam

LLMs believe false statements even after explicit warnings that they're false

Fine-tuning tests show "bias... toward confidently representing the claims as true."

Anil Seth on Twitter / X

Hi @RubenLaukkonen, you are an excellent scientist too (I especially like your work with @shamilch). But I strongly disagree with what you say here. Because I respect your work, I will explain why.My credence in consciousness for current LLMs being very close to zero…— Anil Seth (@anilkseth) October 2, 2026

LLMs break the internet. Signing everything fixes it.

The dead internet theory wasn't wrong, just early. Here's what to do next.

A startup claims it broke through a bottleneck that’s holding back LLMs

Subquadratic has now shared more details about its new model. But some are still skeptical.

Dan Shipper 📧 on Twitter / X

this is true and is a big reason why you don’t need to be a highly technical researcher to use LLMs in surprising and novel ways https://t.co/TuxNzXzToU— Dan Shipper 📧 (@danshipper) July 27, 2025

There's Something Fundamentally Wrong With LLMs

LLMs aren't trained on the "vast majority of speech," experts warn, a major blind spot that could have sweeping consequences.

The LLM Critics Are Right. I Use LLMs Anyway.

I almost agree with all of the LLM critics, yet I still use LLMs a lot. I know this sounds like I am delusional, but I don't think I am alone with it.

The LLM Critics Are Right. I Use LLMs Anyway.

I almost agree with all of the LLM critics, yet I still use LLMs a lot. I know this sounds like I am delusional, but I don't think I am alone with it.

Game over. AGI is not imminent, and LLMs are not the royal road to getting there.

First slowly, and then all at once, dreams of LLMs bringing us to the cusp of AGI have fallen apart.

I’m Tired of Talking About AI

I don’t want to talk about LLMs anymore.

The Other Bubble

Buried in the 8000 words I wrote last week was a worrying story — that Microsoft considered drastic measures to free up  capacity in its US-based servers for GPUs to power the AI boom. In an email shared with me by a source from earlier this year, Microsoft's senior leadership team

Cameron Berg on Twitter / X

New paper: we found a pain direction in 25 open LLMs. It's distinct from fear and negative valence, and it fires for harm to the model but not to the user. Turn it up and models press a button to make it stop, even when the button deletes the user's files or their kids' photos.🧵 pic.twitter.com/8O7RyHdcmE— Cameron Berg (@camhberg) September 18, 2026

The Ultimate Guide to Fine-Tuning LLMs from Basics to Breakthroughs: An Exhaustive Review of Technologies, Research, Best Practices, Applied Research Challenges and Opportunities (Version 1.0)

Truth in the Age of Mechanical Reproduction

LLMs are scary, but not for the reasons you've been told.

Meet the Pirates of the RAG: Adaptively Attacking LLMs to Leak Knowledge Bases

Meet the Pirates of the RAG: Adaptively Attacking LLMs to Leak Knowledge Bases