Begun the document wars have No, seriously. www.scmp.com/economy/chin...

I wrote up my notes on creating Claude Code plugins This is a very powerful system indeed. Still got some rough edges around installation/removal

Reposting this, because I found this short article an encouraging model of how one academic discipline is proactively engaging AI and reshaping it for disciplinary needs.

Interview with the CEO of Google Cloud on AI and agents. He gives yet another definition of agents but also takes about “Office”, “O365” and “Outlook” while not mentioning Google Workspace once. Workspace is part of his empire.

I have two free hours on a train, so I'm using it to learn about SAP ABAP via Claude. I must have been bad in a previous life but I'm quite enjoying the crazy mix of COBOL, SQL and Pascal syntax and random German v English words

The tl;dr of this article is that Peter Theil is fucking weird. But also we should take it seriously - he has an driving philosophy and is in a position to make it happen The antichrist is the rationale for creating the ultimate surveillance state, rooted in elite superiority

"Make Argentina Gold Again" - this post only sets out co-incidence (by it's own admission) and I'm not totally sold, but it is quite the set of links between Trumpworld, Tether and Milei Even if this is 2+2=5, it's a good reminder that Tether is both sketchy af and deeply linked to Trumpworld

Okay so let's talk through this: 1. He's 80 years old. He doesn't know how to embed a video. So this was likely a staffer. 2. Why would a staffer post this? It's likely just chum for his base, since traditional media has given up on even covering him. 3. That's arguably worse than him posting it.

The logic is irresistible - med beds must exist because otherwise q breaks down. Lit crit as the new essential life skill and AI generated slop for ever

How not to illustrate the similarity of bone structures between humans & horses.

I'm at an ERP event for a big German ERP provider and their UK partners. ERP is not exciting, but the keynote went big on "hyper", "lightspeed", "10x automation" and that we should "calculate our hyperdrive jump". The presenter was wearing an astronaut suite. Argh!

Some really, really good news on solar. Solar and batteries are getting very cheap indeed, and can be built at massive scale.

This is real and I need to have a lie down

We all need to be having these conversations.

I’m as centrist dad as it gets. Big fan of both sides-ism. I’m not voting for this bunch again until they have a complete change of leadership and a backbone implant

More things to lighten the day - "If Silksong was made by Ubisoft" youtu.be/j76qWVIW_FQ?...

Tom Stafford (@[email protected]) on distributed peer review I didn't know that astronomers already use this to allocate time on telescopes

In a grim week, this really did make me laugh.

This is a nice and brief summary of the problems Oracle has coming. They've booked revenue (or at least said it's coming) without the associated costs

The “Twitter is more influential than when it was more mainstream” frog boiling radicalisation argument feels pretty convincing. Elite capture!

I think that if you are trying to understand the Anthropic book settlement, and you are not doing it in light of the (extremely) failed Google book settlement from 2011, you will never be able to understand what is happening.

there's more friction. The classic 8-10 pp take-home research paper does overlap with capabilities of language models. People have proposed dozens of patches for this, but I'm skeptical that patches will hold. It's not just teaching; a deeper convo is required about knowledge generation in the 21c.

I'm teaching a 100-level literature class this semester that is totally device-free; we read things on paper and write on paper, in person. Works great when the goal is for students to internalize existing knowledge. However, when students get to be juniors and are working on original projects, +

This thread and the discussion around it are thoughtful about AI and teaching

This is a really good article. The (lack of a) link between grade inflation and mental health and the resulting changes in behaviour around how to get a job

I'm writing a chatbot as an experiment. Here's Claude Code debugging why the tool calling isn't working Spoiler - this was not the reason!

2022 the best AI I had access to failed my Internet Protocols exam badly. This year it gets 100% and helps tighten up loose wording in questions and spots where I messed a question up. I can automatically catch careless slips in exam setting. First time I saw it get subnetting right.

The overreach in the other direction gets under way... Sigh.

Wild thread about the Titan sub that imploded

gpt-5 in cursor does some odd things that neither the default model in Cursor or Claude Code don't. It's very keen on running commands via zsh /bin/zsh -lc 'uv run pyrefly check | cat' The default model correctly calls uv run pyrefly check directly

gpt-5 is very, very good at code based on some quick testing. I don't seem to have API access in the UK but access via Cursor works. It's just a shame Cursor isn't as nice as Claude Code, but it is a lot better than last time I looked at it

The chart crimes will continue until morale improves #gpt5

Research on the scale of academic fraud/paper mills. It's not a cheery read. The problem is large, systemic and growing

On reflection, the difference is that in 2023 people were still feeling confident about responses like “it can’t cite sources, so this won’t affect my class.” As that confidence has faded, we’re seeing a subset of faculty coalesce around a strict zero tolerance position.

I still think we’re getting beyond denial—but wow, skimming the QTs on this you would not know it. This reads to me as maximally cautious guidance. About as negative about AI as you can get without saying “zero tolerance for any use at all.” It’s getting flamed for not being zero tolerance.

I'm reviewing SAP contractor CVs. There's a lot of "15 years+ of ECC to S4/HANA migration experience". S4/HANA came out in 2015

Anthropic cut OpenAI out from Claude Code 😂

Numbers from Google on tokens. Doubling every month. 1,000tn in June.

Incredibly instructive reporting on what appears well on its way to becoming the biggest, most concentrated, most energy- and resource-intensive, and most speculative wave of capital investment ever. ig.ft.com/ai-data-cent...

wtf?! It never, ever occured to me that wikipedia might get caught up in this

New development in the Q Developer extension backdoor: An open question was, how did the threat actor get write permissions to the repo? Turns out they stole a token via this technique. aws.amazon.com/security/sec...

Oh my god

official DHS account posting fourteen word long tweets with the H's randomly capitalized groyper occupied government

I just tried the card route, as I'm ok with this flow (*). Tried two different cards. Both declined. (*) I'm not ok with the principle, but needs must

Worth noting this describes the credit card route but not the face scan route, which is via a different provider. (And IME is quicker)

It's here. Sigh. Age ID for DMs

This is *wild* and a very bad look on @awscloud.bsky.social part. A new github account submits a PR to Amazon Q (their LLM coding tool) with a prompt that is "lol, delete it all". It gets merged and pushed out to customers and the response is to silently pull it. Great reporting by @404media.co