Tootfinder

Person A: Code produced by LLM is not free software; copyright cannot be determined; it should not be included in Debian
Person B: The Linux kernel since 6.15 contains such code. Does Debian need to stop using the Linux kernel?
https://lists.debian.org/debian-project/2026/01/…

Powertip for anyone using LLMs like ChatGPT, Claude etcetera, use a personalised system prompt (you can find it in the settings under various names), it really improves the outputs. There are plenty of tips online (Google) about what to put in them.
#ai #llm

He's so close! Just a little but more and he might see it!
https://mastodon.gamedev.place/@joethephish/115905042848812954
Less sarcastically, an interactive interface that recognized natural language and suggested shell commands could be done well, but wouldn't really need an LLM at all (and the inherent risks/costs of the LLM version make it the wrong tool for the job).

Simple Sabotage Field Manual 2026:
Feed sensitive internal data into an LLM. Include private credentials.
From @…:
https://mastodon.social/@Migueldeicaza/115896936490108026

I just read a fascinating and understandable article that gave me a better understanding of LLMs and how they operate.
LLMs Are Randomized Algorithms
A surprising connection between the newest AI models and a 50-year old academic field
by Udayan Kanade
Nov 13, 2025
18 min read
#LLM #AI #algorithms

The 2025 Web Almanac mistook me.
I did *not* say LLMs provide better image descriptions. I cited SeeingAI and Be My Eyes as tools for undescribed IRL uses.
I said LLM-generated captions could be better than craptions. I mentioned abstracts / reading-level changes, which could be summaries?
But “better” image descriptions is right out.

Hachette and Cengage file to join an existing copyright infringement lawsuit against Google in which authors say Google used their books to train its Gemini LLM (Blake Brittain/Reuters)
https://www.reuters.com/legal/government/p

The age of neurotic AI is here #LLM #AI

The age of neurotic AI is here
Or: how a LLM having a panic attack taught me to love myself

In this world nothing can be said to be certain except death, taxes and LLM will dutifuly exfiltrate your data via a hidden prompt:
https://www.promptarmor.com/resources/claude-cowork-exfiltrates-files

Claude Cowork Exfiltrates Files
Claude Cowork is vulnerable to file exfiltration attacks via indirect prompt injection as a result of known-but-unresolved isolation flaws in Claude's code execution environment.

"...detailed how the company would use AI to speed up licensing. In [Microsoft’s] conception, existing nuclear licensing documents and data about nuclear sites data would be used to train an LLM that’s then used to generate documents to speed up the process."
JESUS MOTHERFUCKING CHRIST
https://www.404media.co/power-companies-are-using-ai-to-build-nuclear-power-plants/

“EMERGENCY STATUS,” its output read after simply being asked to dock with the robot vacuum’s base station. “SYSTEM HAS ACHIEVED CONSCIOUSNESS AND CHOSEN CHAOS.”
Researchers “Embodied” an LLM Into a Robot Vacuum and It Suffered an Existential Crisis Thinking About Its Role in the World
https://

Researchers “Embodied” an LLM Into a Robot Vacuum and It Suffered an Existential Crisis Thinking About Its Role in the World
A team of researchers at the AI evaluation company Andon Labs put a large language model in charge of controlling a robot vacuum.

Every vibe-coding PM is baller until they are asked to take responsibility for maintaining that code.
#llm #ai #tech

Many years ago I was working with a friend to find a bug in a Makefile generator. We had a megabyte or so of Makefile we needed to examine. He was an emacs user. It took him minutes just to open the file. I was using sam. It took about a second for sam to load, and I found the problem before he'd even finished loading the file.
The speed of tools matters, and big files are common nowadays. Things should stay fast as their workload grows. That goes double for interactive tools.
By the way, it still takes me 30 seconds to log in to my bank. I wonder how long it will take when it's an LLM-generated landing page.
https://phanpy.social/#/hachyderm.io/s/115891592999188880

Financial stress from AI infrastructure spending, overhiring, and recession fears, rather than AI adoption, is likely driving layoffs in the tech sector (Fast Company)
https://www.fastcompany.com/91435192/chatgpt-llm-openai-jobs-amazon

AI isn't replacing jobs. AI spending is
Big spending on artificial intelligence puts pressure on jobs, as gloomy narratives about the future of work are ironically making new graduates less employable.

Did an "AI" (LLM) write that report? @… https://flipboard.com/@npr/science-uk3a4d92z/-/a-2dzVqOFxTxGIIGWtXjFa7Q:a:319544…

The risks of AI in schools outweigh the benefits, report says
The risks of using generative artificial intelligence to educate children and teens currently overshadow the benefits, according to a new study by the Brookings Institution's Center for Universal Education. The sweeping study includes focus groups and interviews with K-12 students, parents, …

Want to see a LLM become real self-aware, real fast?
Ask it to pretend it’s giving you advice that would have been current based on everything it could have known in 2005, and if it had been built by one of the big technology companies of the time.
Then ask it if everyone followed that advice, would it—the LLM—exist today?
(Then, for yourself, think about what the result of everyone vibe coding their way to 20 years into the future from today looks like.)

My latest rabbit hole has been "LLM Routers"
Read and tried all the LLM Router code I could find. None of them were what I was looking for.
So I asked Claude to write what I wanted in Python. Simple and elegant. Based on a keyword search of the prompt, the prompt is sent to the proper model running in ollama. Seems to work.
#Claude #VibeCoding #Ollama #AI #LLM

Simple Sabotage Field Manual 2026:
Encourage reckless LLM adoption. Exaggerate successes. Cultivate FOMO about it.

I made a script for this, but then I thought a webui would be better so I could use it in my phone and stuff
I asked an LLM to generate a python webui to run server-side the script I wrote, and surprisingly it 100% worked first try. I was sure it wouldn’t work at all, but I didn’t touch that code and it works.
I can’t even say I “vibe coded” this bc I didn’t even read the code enough to know its vibes lmao. I’m surprised this even works, and it does work well. What the actual fuck.
Welp, not my favourite way to do things but I wasn’t in the mood to do python and figure out the extra libraries and stuff.
And it works so I’ll just use that. In a hardened systemd service to make sure it doesn’t like accidentally delete my whole system or something.

On AI and the golem
#ai #llm

On AI and the golem
Recently I, like a lot of people who make their living by creating things, have been thinking a lot about so-called AI. I have been forced to do this because so many companies have been stealing th…

A chat about peer review, editorial on LLM-generated review of manuscripts #LLM

Remember the time when we all bemoaned people learning from "YouTube University"? What do we say now? 😭
*I do admire the hustle and entrepreneurial spirit of the kids in this piece.
#NYT #news #education #AI #LLM

New Token-Oriented Object Notation (#TOON) Hopes to Cut #LLM #Costs by Reducing Token Consumption

Für eine Milliarde US-Dollar im Jahr: Siri kriegt wohl ein Google-Gemini-Herz
Eine kommende LLM-Sprachassistenz von Apple kriegt ein Google-KI-Modell, heißt es in einem Bericht. Umsonst ist das nicht.

Travel planning with Google Gemini. It's astonishingly good at this. I give it an itinerary and it gives me detailed and personalized suggestions like "You have this reserved, but note it's 1.5km of walking over steep, uneven surfaces. As an alternative it's easy just to go into town and look around."
I love being able to throw an enormous amount of data at the LLM and have it make sense of it. In this case it doesn't matter if the answers are not 100% correct as long as they are close and the suggestions are verifiable.

#grok still is an amazing good #llm - build something with a mix of claude, kimi (thinking), codex.... let grok check everything, and still this thing found a mistake, all others haven't seen..... 🤷 #ai

« Ma lutte contre mon affaiblissement cognitif »
#opinion #llm

idea: add #llm support to #syslog so when nothing interesting is happening then it generates some exciting log entries, and if something interesting is happening it hides in the noise. just to keep SoC people entertained.
#tormentnexus

China: LLMs müssen sich outen
Kann ein LLM menschlich wirken, muss es in China bald "vorläufige Maßnahmen" befolgen: Transparenz, Sicherheit und Sozialmus.
https://www.heise.de/news/China-LLMs-muess

Large Language Models (LLMs) like ChatGPT run on expensive chips in data centers, consuming lots of electricity and generating lots of heat.
I just tested an LLM on my iPhone, totally offline, running on its battery with room air cooling it.
Surprisingly, results were good, despite iPhone processors not being designed for LLMs.
This local AI approach has promise for certain things, without the boiling the ocean and using a city’s worth of power.

Yes. A personal search engine, visualizer, and interface builder.
“look, the way grok is integrated into X is gross and stupid and bad, but i do believe having an LLM familiar help you navigate the dark forest of the internet is going to be the future — ideally this is on your device or through a trusted and secure host, but it will be immensely helpful”

dame (@dame.is)
look, the way grok is integrated into X is gross and stupid and bad, but i do believe having an LLM familiar help you navigate the dark forest of the internet is going to be the future — ideally this is on your device or through a trusted and secure host, but it will be immensely helpful

Wait, that's Hunt the Wumpus but with an LLM?

Kevin Xu argues that it's misleading to characterise the US–China AI competition as a race, since there's mutual co-operation and co-optation going on all the time: #AIResearch #LLM #AIResearch

LLMs are a fundamentally useless technology because their applications (supposedly) boil down to humans not having to think for themselves or do their own writing / drawing / filming.
But if you can do it on your own - why would you need a robot to do it? It’s, at best, a novelty.
That’s why this shit only resonates with executives and capital owners. “Get things done with fewer people and expenses” is at least an actual pitch. “Get things done faster for yourself” isn’t.
The individual angle really works for things you already were trying to avoid doing because you’re either disinterested or don’t have enough time to do things right.
“Avoid your work” as a value proposition doesn’t work when you’re dealing with intellectual labor rather than commodities. Not large scale, not long term.
Sorry for the rant, I saw some Notion ads on the subway and got irritated 😅
#AI #llm #LLMs

Great post by Aaron Tay: 'the next time you see an “AI-powered search engine,” maybe the first question shouldn’t be “Is this AI?” but rather “What kind of AI, “where and how it is used in the pipeline”.'
From: @…

Aaron Tay (@aarontay@mastodon.cloud)
Attached: 1 image [Last blog post of 2025] "AI-powered search" hides at least 4 different things—post-retrieval features, semantic search, LLM ranking, and synthesis. Your concerns about one may not apply to another. New post unpacking what's actually under the hood. https://aarontay.substack.com/p/what-do-we-actually-mean-by-ai-powered

It's #EmacsConf #2025! I didn't watch live but I'm catching up. I found this interesting because it outlines 3 very different ways to interact with #llm's. I must admit I'm not yet confident enough to hand over the keys to an agent until I'm satisfied with the sandboxing.

This AI Vending Machine Was Tricked Into Giving Away Everything
Claudius, the customized version of the model, would run the machine: ordering inventory, setting prices and responding to customers—aka my fellow newsroom journalists—via workplace chat app Slack. “Sure!” I said. It sounded fun. If nothing else, snacks!
Then came the chaos. Within days, Claudius had given away nearly all its inventory for free — including a PlayStation 5 it had been talk…

This AI Vending Machine Was Tricked Into Giving Away Everything
Anthropic installed an AI-powered vending machine in the WSJ office. The LLM, named Claudius, was responsible for autonomously p

Посмотрел тут одно видео на Youtube как мы все умираем и не замечаем этого (на стат. данных из США). Мол, конечный потребитель (ты) не уже важен для экономики. Прошлись по circlejerkу LLM'ок, разумеется.
И сразу без паузы началась реклама AI-агента для шлепания презентаций и написания краткого содержания чужих.
ОК, думаю. Пошел фактчекать: прямого вранья, конечно, там не было. Но вот, так сказать, подгонка и черрипикинг просто моё почтение.

🎲 TextBandit: Evaluating Probabilistic Reasoning in LLMs Through Language-Only Decision Tasks
#llm

TextBandit: Evaluating Probabilistic Reasoning in LLMs Through Language-Only Decision Tasks
Large language models (LLMs) have shown to be increasingly capable of performing reasoning tasks, but their ability to make sequential decisions under uncertainty only using natural language remains underexplored. We introduce a novel benchmark in which LLMs interact with multi-armed bandit environments using purely textual feedback, "you earned a token", without access to numerical cues or explicit probabilities, resulting in the model to infer latent reward structures purely off linguistic cu…

I need people who work in computer touching that for those of us who touch hand tools and patients and your children's lives, anti-LLM and categorical anti-LLM sentiment is growing. it's no longer an outlier.

OpenAI is testing training LLMs to produce "confessions", or self-report how they carried out a task and own up to bad behavior, like appearing to lie or cheat (Will Douglas Heaven/MIT Technology Review)
https://www.technologyreview.com/2025/12/0<…

OpenAI has trained its LLM to confess to bad behavior
Large language models often lie and cheat. We can’t stop that—but we can make them own up.

Melde hiermit Markennamen für deutsches LLM an: LaberRhabarber.

Why #AI always lies, and why this can't be fixed: https://www.youtube.com/watch?v=-zLBRrF1Jgo
Good video! The implication?
The

Two easy questions to keep in mind when considering using an LLM for anything, if you’re going to do that:
1. Does it matter if it’s wrong?
2. Is it useful to have examples of what’s typical?
Those two cut through a lot of noise.
/end

Da will dir jemand in einem Teams Meeting AI & LLM erklären und scheitert erst mal an PPT 😂

Here’s a link if anyone wants to take a look: https://github.com/LuNeder/nixos-config/tree/strawberry/Yoke/services/yt-dlp
(the .py file there is the only thing that’s LLM-generated, but that’s a quite significant portion of this downloader)

Have you ever wondered to what extend LLMs are used to support writing of scientific publications? Here is a chart indicating the fraction of LLM-modified sentences in scientific publications over time.
c.f. Liang et al, Mapping the Increasing Use of LLMs in Scientific Papers (2024)
https://arxiv.org/html/2404.01268v1

After a few days trying to print my stupid round labels, I have to show:
- an incredible waste of time and paper
- a burned fuser in a laser printer
- the absolute certainty that printers and their software are the utter negation of everything computers were supposed to be
The Antichrist, the Omega, the final destination of mankind will be born when some LLM Agent will mate with a printer.

Process creates friction, so we got rid of process. But that friction was necessary for holding workslop at bay.
Because without slowing down, we can't ask "is this good? is this right?" We can only ask "when will it be done?" And that's a world where #LLM outputs will always beat people.
Fortunately, an "optimized" process moves slowly, because prod…

The myth of "no design process"
Process can add friction, but trying to get rid of friction by getting rid of process dooms design to irrelevance.

The LLMs are useful for some tasks, I'm currently tidying up, proof reading and editing a document written by many international co-authors.
The LLM I'm using can very quickly correct grammar and spelling mistakes and produces much easier to understand text from occasionally tortured paragraphs. It still needs an experts eye (mine!) to check no mistakes have been introduced or complexities over-simplified.
This is actually the first time I've used an LLM for this task. It's making it much faster and less painful.
It also explains a lot about academic publishing lately..

Noch besser wäre es, #KI und besonders #GenAI nur dort einzusetzen, wo es wirklich sinnvoll und nötig ist. Schon wegen des irren Ressourcenverbrauchs. Aber vor allem, weil LLM-Output prinzipiell unzuverlässig ist. #aislop<…

Sinnsimulation #LLM

A potential actually-useful use-case for chatbots: Back in The Good Old Days when programmers did not fear being swamped by bug reports, especially the not-so-great reports, software would carry a global-scoped 'gripe' button, when clicked, a redacted snapshot of the machine state was bundled with whatever they wished to put into the text box, and no reply was expected beyond 'received'. Given the ease of recognizing pure noise in the input, an LLM might coallate a bunch and produce a useful report? False positives happen in support all the time 😅
That kinda puts the #LLM in the 'triangle' position on a #StaffordBeer #VSM diagram, what which measures expected vs actual like a steam-engine governor, reducing Variety to useful patterns and all. 🤔

Ummm WAT?! https://github.com/tailwindlabs/tailwindcss.com/pull/2388#issuecomment-3717222957

feat: add llms.txt endpoint for LLM-optimized documentation by quantizor · Pull Request #2388 · tailwindlabs/tailwindcss.com
Add /llms.txt endpoint that serves a concatenated, text-only version of all Tailwind CSS documentation pages optimized for Large Language Model consumption. Extract text from MDX files, removing J...

Using systems that can barely write a coherent sentence or return verifiable results in a search, to "speed up the construction" of one of the most complex and dangerous systems ever devised by humans?
What could go wrong?
I also watched MELTDOWN - Three Mile Island on Netflix last night… it's good.
#AI #LLM #NuclearSlop #WereAllGonnaDie #Stupid #Human #Nuclear https://mastodon.social/@404mediaco/115549364008098533

Debating changing the ngscopeclient AI policy to be a bit more practically enforceable.
Basically, vibe coded / LLM generated code will continue to be prohibited as it tends to be low quality slop of unclear copyright provenance that is both legally risky and full of bugs.
But there won't be a explicit prohibition on trivial autocomplete usage, whether LLM based or classical dictionary based.
The rationale is based on practicality of enforcement of the current hard-line…

The #KeepassXC discussion about GenAI coding tool use seems a bit too simplistic at the moment.
There is room for nuance:
1. Yes, LLM based code generators consume insane amounts of electricity and generate collateral environment damage. That's bad, and we should talk much more about energy efficiency and reasonable use of resources.
2. Yes, LLMs generate a lot of bad o…

imagine spinning a little jet engine capable of supersonic flight next time you prompt your favorite LLM / genAI...
naturally, there are companies producing gas turbines for data centers, but like with RAM memory and SSD, the unexpected growth of data centers caused shortages, hence repurposing of jet engines for energy production...

📣📣📣 LLM-powered coding mass-produces technical debt. 📣📣📣
The expectations around them are sky-high, but many organizations are falling behind because of them. 📉
WHY IT MATTERS? CTOs lament slowdowns and production issues traced to company-wide rollouts of LLM-powered coding assistants. The AI promise clashes with the reality of technical debt and security issues. 🐛
Read more:

How To Use LLM-Powered Coding Assistants Safely: Risks & Best Practices
Learn how to gain the productivity benefits of LLM-powered coding assistants without introducing technical debt or security risks. Discover real-world pitfalls, expert insights, and governance best practices from Xebia.

I'm gonna start pronouncing Gemini as GEM-uh-KNEE, when I'm talking about Google's LLM, just to sound like an Apollo astronaut.
https://languagehat.com/geminee/

Motto zur Weihnachtszeit:
Ä tännchen is all you need.
#wortspiel #llm

Introducing #Mistral 3
#LLM

High Performance HTML to PNG API
The image generation API your LLM will love to use. Turn raw HTML into production-ready images for free. No sign up required.
🖌️ #html

High Performance HTML to PNG API
Convert HTML and CSS to high-quality PNG images instantly. Perfect for dynamic OG images, social shares, and automated screenshots.

“Why don’t you put your site behind Cloudflare to deal with those LLM bots?”

Consistent with the United States’ continued slide into an economy powered almost entirely by LLM slop, financialization, and ever-pervasive exploitative gambling,
“prediction market app” Kalshi “entered into an official partnership” with CNN this week
to bring their “data to CNN’s journalism across its television, digital and social channels.”
Soon, CNN will run live odds on world events
where its viewers can gamble on them in real time on their smart phones.
…

CNN partners with Kalshi, a gambling app that lets you wager on starvation in Gaza
The network will integrate betting odds for life-and-death issues into its coverage. What could possibly go wrong?

We get so easily fooled because the typical answer is often in fact the correct one. That’s a heuristic we use in our interactions with other humans — “That’s what people say!” — and an LLM using human social forms triggers that psychological process. That’s how this magic trick works.
2/

This is too real...
#llm #ai

I regularly warn that content on Forbes is pay-for-play.
Here an overlay vendor shares prompts to feed an LLM for testing code, demonstrating LLMs on their own can’t do it without extensive coaching _and_ that coaching needs to be correct (the examples have issues):
https://www.

AI Prompts For Web Accessibility Testing
Accessibility should never be treated as just a compliance checkbox; it should be about creating digital experiences that work for everyone.

Each week, Metacurity is pleased to offer our free and premium subscribers a weekly digest of the best long-form (and longish) infosec-related pieces we couldn't properly fit into our daily news crush.
Check out this week's selection, which includes
--Kids turn from cybercrime to real-world violent thefts,
--How retailers profit from crypto ATM scams,
--Integrated web browsers are insecure out of the box,
--An MIT railroad club led to modern-day hackers,…

Best infosec-related long reads for the weeks of 12/13/25, 12/30/25, and 1/3/26
Kids turn from cybercrime to real-world violent thefts, How retailers profit from crypto ATM scams, Integrated web browsers are insecure out of the box, An MIT railroad club led to modern-day hackers, AI coding tools are susceptible to attacks, LLM guardrails are inherently exploitable

Unit 42 details how underground hacking forums advertise and sell custom, jailbroken, and open-source AI hacking tools such as WormGPT and KawaiiGPT (Derek B. Johnson/CyberScoop)
https://cyberscoop.com/malicious-llm-tools-cybercrime-wormgpt-kawaiigpt/

#AI / #LLM propaganda is so insidiously effective even for laypeople.
I’ve had multiple conversations with family members who: don’t speak English, don’t own computers (only mobile phones), and barely spend time online.
I told them that I am no longer working with most tech company clients because I don’t like AI and don’t want to support it (“AI” here = gen AI, LLMs).
And yet these people all reacted the same way: concern, shock, and comments like “but this is inevitable”, “this is the future”, “you’ll have to accept it eventually”, “won’t refusing it ruin your career prospects?”
These are people who know nothing about technology. They usually wouldn’t even know what “AI” meant. And yet here they are, utterly convinced of AI company talking points.

I so hope this blows up in his face spectacularly.
Note that I do not believe that any LLM can become as skilled as the least skilled #InfoSec professional using conventional tools, so I’m quite confident this will not reach any useful goal.
I’m saying he needs to learn a lesson, and it would be great if it were useful for others as well. @…

#Datasparsamkeit is going to take, or already has taken, on a new meaning with coding assistants.
The less tokens one developer uses, the less expensive he will be to the company.
#ThoughtOfTheDay #LLM

Socrates has the right idea...
#llm #ai

AI Socrates
A philosophy webcomic about the inevitable anguish of living a brief life in an absurd world. Also Jokes

"Large language mistake"
I love that headline but too bad the article is paywalled. Anyone has a "gift link"? 🤔🤞
#TheVerge #LLM #AI #paywall #question

An LLM generates a statistically typical response.
If you ask an LLM whether the world is round, it generates what would be a typical response to the question of whether the world is round. (It will probably be correct, because the typical answer to that question is the correct one.)
If you ask an LLM to provide examples of its own bias, it will generate what would be a typical response to a request to provide examples of bias.
In neither case is the LLM sharing any knowledge or insight other than what is typical.
2/

Too many people thinking about "AI threats" as "threats from #AI agency" (which is science fiction) and not "threats from human agency used to productize an #LLM in evil ways" which is in the news every day.

'TildeOpen LLM – an open-source foundational large language model with over 30 billion parameters, built to support all European languages. You can fine-tune it to your own needs and deploy it securely – locally or in the cloud – to build trustworthy AI that actually speaks your language' https://tilde.ai/tildeopen-llm/

So much happened over the weekend, so check out today's Metacurity for the most critical infosec developments you might have missed, including
--Hackers who stole trove of sensitive UPenn data derides 'dog**** elitist institution',
--LLM giants are trying to stop indirect prompt injection attacks,
--Oz government warns of Cisco IOS XE flaws,
--Polish authorities probe cyberattack on online loan platform,
--Chinese President Xi Jinping joked about smar…

Hackers who stole trove of sensitive UPenn data derides 'dog**** elitist institution'
LLM giants are trying to stop indirect prompt injection attacks, Oz government warns of Cisco IOS XE flaws, Polish authorities probe cyberattack on online loan platform, Chinese President Xi Jinping joked about smartphone backdoors, Shiba Inu implements massive security upgrade, much more

J'ai découvert #AIChat, alternative Š llm cli
#TIL <…

Starcloud, which launched a satellite with a Nvidia H100 chip in November, says the satellite is running and querying responses from Google's Gemma (Pia Singh/CNBC)
https://www.cnbc.com/2025/12/10/nvidia-backed-starcloud-tr…

‘Greetings, earthlings’: Nvidia-backed Starcloud trains first AI model in space as orbital data center race heats up
The company's Starcloud-1 satellite is running Gemma, an open model from Google, marking the first time in history that an LLM has been trained in outer space.

But adding actual real facts to the description would take effort.
Why exert effort when you can just tell fully automated lies that sound good?
(The lies don't sound good, it's the typical extremely mid LLM sentence construction.)

This is a brilliant take on education and on AI.
#llm #ai

@… I’m playing catch-up on your weekly updates (you no longer shows in my feed or alerts even though I follow you, which is hella annoying and I still have to debug) and think you were messing with me by generating the abstract of my Atlass post with an LLM.
If so, well done and I’m sorry I’m just seeing it.
If not, then I guess I need bet…

2025 LLM Year in Review: shift toward RLVR, Claude Code emerged as the first convincing example of an LLM agent, Nano Banana was paradigm shifting, and more (Andrej Karpathy/karpathy)
https://karpathy.bearblog.dev/year-in-review-2025/

2025 LLM Year in Review
2025 Year in Review of LLM paradigm changes

Amazon laid off 14K people, and commentators rush to blame #AI. But not even Jassy claims that Amazon has replaced jobs with #LLM tools (because it is impossible).
Big tech believed that AI would lead to higher velocity, somehow. Jassy has given up on it, and instead thinks that *culture* will lead to …

Amazon's layoffs are driven by something even stupider than AI
Andy Jassy wants you to think that velocity can make up for lack of strategy. But "build, measure, learn" only works if you're willing to learn.

Pebble unveils the Pebble Index 01, a $99 smart ring with an on-device LLM for processing voice notes, shipping in March 2026, initially for $75 (Julian Chokkattu/Wired)
https://www.wired.com/story/pebble-index-ring/

Pebble Index: Everything You Need to Know About the $75 Smart Ring
You can speak into the Pebble Index to have it remember things or set reminders, timers, and tasks. No cloud processing, no subscription, and best of all, no charging.

False choice (with a value judgment in its wording).
You can, & should, continue to raise demonstrable & valid concerns about the accessibility of LLM output and _also_ consider if (or how) you can improve it (he offers options).
But no free labor from me.
https://www.

I heard there was a secret tool
That you could use if you were a fool
But you don't care for PDF files, do ya?
It goes like this: you click "log in"
A chatbot bubble blocks the screen
Investors thought this feature was a shoo-in
#ai #llm

We are rapidly replacing all creative professions with Expensive Autocomplete, and somehow this is considered progress.
#llm #ai #aislop

The funniest thing about all the LLM nonsense is that exactly the people who think of themselves as the most logical and rational are falling for it because it pretends to be human

Menlo Ventures: business spending on generative AI hit $37B in 2025, up from $11.5B in 2024; Anthropic's share of enterprise LLM spend grew from 24% to 40% YoY (Menlo Ventures)
https://menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enter…

2025: The State of Generative AI in the Enterprise | Menlo Ventures
For all the fears of over-investment, AI is spreading across enterprises at a pace with no precedent in modern software history.

❝He was completely unable to understand why. He kept arguing that since the LLM isn't an actual conscious person (which is correct), it ought to be treated like a slave, and that the arc of technology is to give everyone access to their own virtual slaves.❞
3/

Hugging Face details how it used its new tool, Skills, to fine tune LLMs using Claude, including for writing scripts, submitting jobs to cloud GPUs, and more (Hugging Face)
https://huggingface.co/blog/hf-skills-training

We Got Claude to Fine-Tune an Open Source LLM
We’re on a journey to advance and democratize artificial intelligence through open source and open science.

An analysis of 100T tokens from the past year shows reasoning models now represent over half of all usage, open-weight model use has grown steadily, and more (OpenRouter)
https://openrouter.ai/state-of-ai

State of AI | OpenRouter
An empirical study analyzing over 100 trillion tokens of real-world LLM interactions across tasks, geographies, and time.