(Someone who barely has a mind) large language models are mind blowing
August 9, 2025 at 9:42 PM
(Someone who barely has a mind) large language models are mind blowing
If you have zero education, but learn how to ask AI models the right questions , in many jobs you will be able to outperform someone with an advanced degree, but who is unwilling to use Large Language Models.
Just takes a smartphone, curiosity to experiment and a mindset to learn.
Just takes a smartphone, curiosity to experiment and a mindset to learn.
February 17, 2025 at 5:59 PM
If you have zero education, but learn how to ask AI models the right questions , in many jobs you will be able to outperform someone with an advanced degree, but who is unwilling to use Large Language Models.
Just takes a smartphone, curiosity to experiment and a mindset to learn.
Just takes a smartphone, curiosity to experiment and a mindset to learn.
Print dictionaries: the original large language models
11/18/25
11/18/25
November 17, 2025 at 2:41 PM
Print dictionaries: the original large language models
11/18/25
11/18/25
Journalists stop ascribing agency to large language models 2026 challenge
September 24, 2026 at 1:37 AM
Journalists stop ascribing agency to large language models 2026 challenge
Is this how you think science or generative language models actually work
July 20, 2025 at 9:06 AM
Is this how you think science or generative language models actually work
You've been talking to models? Niiiiiice. Wait, large language?
February 16, 2026 at 9:33 PM
You've been talking to models? Niiiiiice. Wait, large language?
language models of carriage
December 30, 2025 at 5:35 PM
language models of carriage
Last week @theverge.com published my essay exploring the limitations of large-language models. This week, that same essay is cited by a federal judge in Michigan to distinguish the process of human reasoning from what these models do. Very, very gratifying.
h/t @robertfreundlaw.bsky.social
holy shit, an accurate legal critique of LLMs. LLMs don't reason because they're just stitching together plausible-looking sentences indifferent to the content
holy shit, an accurate legal critique of LLMs. LLMs don't reason because they're just stitching together plausible-looking sentences indifferent to the content
December 3, 2025 at 9:04 PM
Last week @theverge.com published my essay exploring the limitations of large-language models. This week, that same essay is cited by a federal judge in Michigan to distinguish the process of human reasoning from what these models do. Very, very gratifying.
Large language theft models (LLTMs)
September 22, 2026 at 9:39 PM
Large language theft models (LLTMs)
"In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limits."
www.computerworld.com/article/4059...
www.computerworld.com/article/4059...
OpenAI admits AI hallucinations are mathematically inevitable, not just engineering flaws
In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limi...
www.computerworld.com
September 21, 2025 at 12:50 PM
"In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limits."
www.computerworld.com/article/4059...
www.computerworld.com/article/4059...
This morning I read that DeepSeek used training data generated from other language models which has companies like OpenAI fuming, and I can't stop thinking about how there's a lesson in there
January 28, 2025 at 12:08 AM
This morning I read that DeepSeek used training data generated from other language models which has companies like OpenAI fuming, and I can't stop thinking about how there's a lesson in there
Surrendering our ability to think and do research to large language models is stupid as shit
Formerly all of these models were useless for my work because 1) hallucinations made them dangerous and you couldn’t trace the origins of a “fact”. Also they weren’t current. But you can now ask for citations for everything with links as it crawls the web and it’s an insanely powerful search tool.
February 19, 2025 at 3:14 PM
Surrendering our ability to think and do research to large language models is stupid as shit
Small Language Models.
June 24, 2026 at 10:07 PM
Small Language Models.
Medically adapted foundation models (think Med-*) turn out to be more hot air than hot stuff. Correcting for fatal flaws in evaluation, the current crop are no better on balance than generic foundation models, even on the very tasks for which benefits are claimed.
arxiv.org/abs/2411.04118
arxiv.org/abs/2411.04118
Medical Adaptation of Large Language and Vision-Language Models: Are We Making Progress?
Several recent works seek to develop foundation models specifically for medical applications, adapting general-purpose large language models (LLMs) and vision-language models (VLMs) via continued pret...
arxiv.org
November 26, 2024 at 6:12 PM
Medically adapted foundation models (think Med-*) turn out to be more hot air than hot stuff. Correcting for fatal flaws in evaluation, the current crop are no better on balance than generic foundation models, even on the very tasks for which benefits are claimed.
arxiv.org/abs/2411.04118
arxiv.org/abs/2411.04118
Do large language models develop "emergent" models of the world? My latest Substack posts explore this claim and more generally the nature of "world models":
LLMs and World Models, Part 1: aiguide.substack.com/p/llms-and-w...
LLMs and World Models, Part 2: aiguide.substack.com/p/llms-and-w...
LLMs and World Models, Part 1: aiguide.substack.com/p/llms-and-w...
LLMs and World Models, Part 2: aiguide.substack.com/p/llms-and-w...
LLMs and World Models, Part 1
How do Large Language Models Make Sense of Their “Worlds”?
aiguide.substack.com
February 13, 2025 at 10:30 PM
Do large language models develop "emergent" models of the world? My latest Substack posts explore this claim and more generally the nature of "world models":
LLMs and World Models, Part 1: aiguide.substack.com/p/llms-and-w...
LLMs and World Models, Part 2: aiguide.substack.com/p/llms-and-w...
LLMs and World Models, Part 1: aiguide.substack.com/p/llms-and-w...
LLMs and World Models, Part 2: aiguide.substack.com/p/llms-and-w...
1. We—@eduede.bsky.social, @mjcrockett.bsky.social, Kevin Gross, and I—have a new preprint on the arXiv today, based on ideas that emerged during an @sfiscience.bsky.social workshop in November 2024:
The unintended consequences of large language models as a labor-augmenting technology in science.
The unintended consequences of large language models as a labor-augmenting technology in science.
The unintended consequences of large language models as a labor-augmenting technology in science
As a labor-augmenting technology, large language models (LLMs) have the potential to accelerate scientific activity across the research pipeline. But even if LLMs perform on par with human experts at ...
arxiv.org
July 21, 2026 at 3:19 PM
1. We—@eduede.bsky.social, @mjcrockett.bsky.social, Kevin Gross, and I—have a new preprint on the arXiv today, based on ideas that emerged during an @sfiscience.bsky.social workshop in November 2024:
The unintended consequences of large language models as a labor-augmenting technology in science.
The unintended consequences of large language models as a labor-augmenting technology in science.
Large Language Models
Are Not
Search Engines!!!!!!!!
Are Not
Search Engines!!!!!!!!
Hi, this is Holden emerging from my Bluesky hiatus to tell everyone that when checking the links in the piece, I used ChatGPT as my search engine (does anyone use Google anymore?). The piece was written by me and edited by humans. So all the links were checked by me and again by our copyeditors.
July 17, 2026 at 11:58 PM
Large Language Models
Are Not
Search Engines!!!!!!!!
Are Not
Search Engines!!!!!!!!
We can't let large language models take the em dash from us — it's too important
February 18, 2026 at 1:52 AM
We can't let large language models take the em dash from us — it's too important
Large language models adopt an authoritarive tone and can sound convincing even when they're full of shit, which coincidentally also describes mediocre white businessmen.
April 9, 2026 at 8:29 AM
Large language models adopt an authoritarive tone and can sound convincing even when they're full of shit, which coincidentally also describes mediocre white businessmen.
How I picture people who are impressed by large language models
October 3, 2024 at 9:07 PM
How I picture people who are impressed by large language models
who called them Large Language Models when Big Unnaturals was right there
‘I felt pure, unconditional love’: the people who marry their AI chatbots
‘I felt pure, unconditional love’: the people who marry their AI chatbots
The users of AI companion app Replika found themselves falling for their digital friends. Until – explains a new podcast – the bots went dark, a user was encouraged to kill Queen Elizabeth II and an update changed everything …
www.theguardian.com
July 12, 2025 at 11:39 AM
who called them Large Language Models when Big Unnaturals was right there
August 18, 2025 at 1:52 AM
Language models can correctly answer questions about their previous intentions.
www.anthropic.com/research/int...
www.anthropic.com/research/int...
Emergent introspective awareness in large language models
Research from Anthropic on the ability of large language models to introspect
www.anthropic.com
October 29, 2025 at 6:21 PM
Language models can correctly answer questions about their previous intentions.
www.anthropic.com/research/int...
www.anthropic.com/research/int...
My university: "Copilot is Microsoft's AI-powered productivity service that uses large language models (LLMs) to help you create content, analyze information, summarize documents, and complete tasks more efficiently."
Microsoft: LOL
Microsoft: LOL
www.microsoft.com/en-us/micros... is it good to build an entire economy and software infrastructure on this
April 1, 2026 at 1:11 AM
My university: "Copilot is Microsoft's AI-powered productivity service that uses large language models (LLMs) to help you create content, analyze information, summarize documents, and complete tasks more efficiently."
Microsoft: LOL
Microsoft: LOL
Or how terribly ai language models do humor.
September 19, 2024 at 12:03 AM
Or how terribly ai language models do humor.