Prof Les Carr

Head of Web, AI & Society at @unisouthampton. Interdisciplinary open science research & science standup comedy. @[email protected]

[Prof Les Carr @lescarr]

I’m a nerdy 59 yr-old guy with a computing and AI job at a university and a side hustle in standup comedy. I’m married to a writer and we have four grownup children, a border collie and a camper van.

[Prof Les Carr @lescarr]

@ItsAndyRyan When your headline quote goes with “amiably-peopled” you know that the rest of the review wasn’t all that.

[Prof Les Carr @lescarr]

@esrtweet @danbri I’m more of a classic /etc/termcap TERM=vt52 kind of a guy.

[Prof Les Carr @lescarr]

When AI shows you what it’s capable of, believe it.

[Prof Les Carr @lescarr]

This quote from an AI paper in Nature tells you everything about the AI field's current attitude to rigour - the fact that our results aren't particularly good now is proof that they are going to be brilliant in a few years!

[Prof Les Carr @lescarr]

I’m watching the Yahoo of the 2020s simultaneously beach itself on the Metaverse and AI.

[Anish Moonka @anishmoonka]

Yesterday Meta told every US employee their computer will now record mouse clicks, keystrokes, and screenshots while they work. All of it goes into training an AI to do their job. In 30 days, 8,000 of these same employees are being laid off.

Reuters got the memo. The wording is the company's own: the recordings will be used to build "AI agents that can perform work tasks autonomously." Reuters also confirmed the May 20 date and the number, 8,000 people, exactly 10% of Meta's global workforce.

Meta is spending $115 to $135 billion on AI infrastructure this year, almost double the $72 billion it spent last year. The entire business only generated $115.8 billion in cash for all of 2025. Meta is now planning to spend more on AI in 2026 than the whole company brings in.

Part of the bill went to a company called Scale AI. Meta paid $14.3 billion for 49% of it last June, mostly to bring in CEO Alexandr Wang. Scale's whole job is to tag and clean the human-written data that AI models learn from. Meta wanted Wang because their old data supply ran dry.

The public internet is almost out of fresh material to feed these models. A group called Epoch AI ran the math and projects the world will burn through its supply of high-quality human-written text on the web somewhere between 2026 and 2032. The industry calls this the "data wall." Google and OpenAI are stuck on the same side of it.

So Meta turned inward, to the most expensive training material money can buy: their own employees doing their own jobs. Mouse movements teach the AI how to move around a screen, click by click. Keystroke logs hand it the exact shortcuts and rhythm an experienced worker uses, the muscle memory of the job. Screenshots show what a finished task should look like. The people being recorded in April are the raw material for the AI that replaces them in May.

This is not just a Meta thing. Amazon laid off 16,000 corporate workers in January. Oracle let go of up to 30,000 of its people, about 18% of the company, on March 31. The cash they saved goes toward $156 billion in AI data centers. The whole pattern across big tech is identical. Record profits and record AI spending, paired with the biggest workforce cuts since the pandemic.

The thing they are building is a software worker that opens the dashboard, reads the numbers, drafts the email, books the meeting, and never needs a coffee break. The training data for that worker is a senior Meta employee doing all of that, on Meta's payroll, one month before their last day.

[Prof Les Carr @lescarr]

@kareem_carr My organisation has perhaps 100 of these workflows mutually interacting in ways that already produce logjams. Where we’re at is the hope that AI can, in some cases, help ease some of the existing logjams.

[Prof Les Carr @lescarr]

@adamconover @amendlocke Mr Conover, I hope you are doing well. Your line manager has signed you up for mandatory “Air Fryer Awareness” training, from which you will be required to demonstrate air fryer ethics and policy compliance capabilities. Deadline for completion of the first 2hr module is Friday.

[Prof Les Carr @lescarr]

Is the problem AI or the streaming industry? Either way, techbros gonna techbro.

[Riley Walz @rtwlz]

made my computer dramatically play BBC news music before every meeting

[Prof Les Carr @lescarr]

This tweet hasn’t aged well.

[The AI Showdown 24x7 @ai_24x7]

AMAZON'S AI AGENTS ARRIVE: THE END OF DEVOPS JOBS?

If AI manages the cloud, is the human sysadmin obsolete? #AINews

Imagine an employee who never sleeps, never makes a typo, and fixes your cloud servers before you even know they're broken. Amazon just built him.

[Prof Les Carr @lescarr]

@amolitor99 @LauraRbnsn Yes this. Glad to make new toast for other people so as to manufacture a happy environment that’s easy for me to live in; really can’t be bothered to make fresh toast for myself because that’s needless extra time and effort.

[Prof Les Carr @lescarr]

All the best AI posts are now coming from the security community and I couldn't be more embarrassed about the future of computer science.

[Lukasz Olejnik @lukOlejnik]

Amazon is holding a mandatory meeting about AI breaking its systems. The official framing is "part of normal business." The briefing note describes a trend of incidents with "high blast radius" caused by "Gen-AI assisted changes" for which "best practices and safeguards are not yet fully established." Translation to human language: we gave AI to engineers and things keep breaking?
The response for now? Junior and mid-level engineers can no longer push AI-assisted code without a senior signing off. AWS spent 13 hours recovering after its own AI coding tool, asked to make some changes, decided instead to delete and recreate the environment (the software equivalent of fixing a leaky tap by knocking down the wall). Amazon called that an "extremely limited event" (the affected tool served customers in mainland China).

[Prof Les Carr @lescarr]

@valigo Although I'm not a mathematician, the instability and unreliability of the AI method reported here is very, very familiar. So much handholding. So much checking. And that's before the verification of the outputs!

[Prof Les Carr @lescarr]

@Aurore_Chery @noteamho @joakial_ Except that they *are* the original English.

[Michael Levin @drmichaellevin]

Final version of this paper with @RichardWatson90 is out!
https://t.co/4tiqmFHPkn
"Machines all the way up and cognition all the way down: Updating the machine metaphor in biology"

(quite a bit different than the original preprint at https://t.co/Yco09orClK).

[Prof Les Carr @lescarr]

AI & semi-structured data FTW!

[Rohan Paul @rohanpaul_ai]

Goldman Sachs is rolling out Anthropic’s AI model to automate accounting and compliance roles completely.

Anthropic engineers have been embedded at Goldman for 6 months, co-developing systems that act like “digital co-workers” for high-volume, process-heavy tasks.

The new setup uses an LLM-based agent that can read large bundles of trade records and policy text, then follow step-by-step rules to decide what to do, what to flag, and what to route for approval.

Goldman says the surprise was that Claude’s capability was not limited to coding, and that the same reasoning style worked for rules-based accounting and compliance work that mixes text, tables, and exceptions.

The bank expects shorter cycle times for client vetting and fewer lingering breaks in trade reconciliation, and slower headcount growth rather than immediate layoffs.

---

cnbc .com/2026/02/06/anthropic-goldman-sachs-ai-model-accounting.html

[Prof Les Carr @lescarr]

The introductory lecture from Computer Science 101, 2026 style. https://t.co/7oS3aOgA6B

[Prof Les Carr @lescarr]

The verdict is in! Leaders at Davos 2026 identified organizational inertia and the failure to redesign job roles as the primary hurdles to scaling AI - it turns out that failures in AI are all very much your fault. https://t.co/L9BeFaUyT5

[Dr Ramine Tinati @TinatiRamine]

Amazing reunion with the fantastic @DameWendyDBE :) lots of catching up (and too much sushi) about the current direction of AGI, safety, and the competing worlds of the east and west! 2026 is going to be a firehose again!