'About 10,000 writers have contributed to Don’t Steal This Book, in which the only content is a list of their names. Copies of the work are being distributed to attenders at the London book fair on Tuesday'. 1/2
Beatles and entourage at tea in Bangor, Gwynedd. Wonder if Mick was invited?
You know what? I dont think i have. This might be the greatest one. Especially if, like me, you have a Welsh-things bias.
Can you think of a better one?
'Of the 244,755 academics recorded [by Hesa], the largest share were in teaching and research roles – 105,630 (43 per cent of the total), which was slightly fewer than in 2023-24.' 1/3
2-year postdoc for someone with a PhD in Classics or Ancient History www.jobs.ac.uk/job/DQS114/p... @hcaatedinburgh.bsky.social
New post, on whether I could get Claude Code to complete a data task that had taken me AGES a decade ago…
kucharski.substack.com/p/how-much-t...
Yep, "it's a great tool as long as you already have a good idea of the answer" isn't ideal...
Two cats on a wall looking at me
End of level bosses
Sunset walk with the crocuses
Want your family to be better informed?
Buy them a subscription to Private Eye from just £10.
➡️ checkout.private-eye.co.uk/SingleItem?I...
I get yelled at for saying this but for many hundreds of years people went to university not to get diplomas or be employable but because immersion in the humanities was considered foundational to a good life, and school must return to its original purpose: the joy of learning.
AI art isn’t real art and AI artists aren’t real artists.
Academic writers often feel uncomfortable but that’s often OK Here’s why patthomson.net/2026/03/08/g...
Black and white cover art for Crowley Time Episode 42: The Food of Lurve, with Odinn Orn Hilmarsson.
La la la, and such! It’s a musical spectacular in the new instalment of my sketch comedy podcast Crowley Time with guest star composer and sound designer Odinn Orn Hilmarsson. Episode 42: The Food of Lurve, available NOW on your podcast app and at crowleytime.com!
@odinnoh.bsky.social
Stop with the “vibe coding” and start doing some algorithm design, foundations of computing, and computational complexity analyses.
Even just doing playful computer science unplugged is better than all the current hyped non-sense.
www.csunplugged.org/en/
If it doesn’t bang on and on about re-writing everything in Rust, it’s not AGI.
Prefer to get your podcasts on YouTube?
Check out the Robot Talk channel and subscribe today! 🤖
www.youtube.com/@RobotTalk
In the land of the vibecoders, the person who still knows what an if statement is, is king.
A cool little Mycena that I photographed in Oxapampa, Perú. Never seen one with this cool stem texture before, no idea which species it could be.
Olympus OM-1 / 90 mm macro lens / 180 images stacked, ISO 200, f/5.
#mushrooms #mycena #fungifriends #olympus
And those were the most plausible use cases. Many, many more have been proposed. For a thorough ("scathing" --Publisher's Weekly) takedown, see THE AI CON, out on May 13!
w/@alexhanna.bsky.social
thecon.ai
4) I need some information and I wish to just pose my question to an all-knowing oracle.
DO NOT DO THIS. Chatbots, even if they could reliably return "the" correct answer, are not a good tech for information access.
buttondown.com/maiht3k/arch...
>>
3) I need to write some code and I can't be bothered to remember the syntax of this particular programming language/boiler plate code is boring, etc.
You can test whether the code works as you intend, but are you really in a position to catch security vulnerabilities?
socket.dev/blog/slopsqu...
>>
2) I need to take a long document and get a summary of it.
DO NOT DO THIS -- unless you don't care whether the summary includes inaccurate information or more importantly excludes important points.
>>
What about the cases where they seem useful? These seem to fall into a few categories:
1) I need to take a set of notes and turn it into a polished document, and I'm in a position to check that it says what I mean.
Ok fine but writing is thinking and you're letting that muscle atrophy.
>>
LLMs are nothing more than models of the distribution of the word forms in their training data, with weights modified by post-training to produce somewhat different distributions. Unless your use case requires a model of a distribution of word forms in text, indeed, they suck and aren't useful.
I was all on board until I got to: "In the longer term, though, making students do all of these things by themself may be the 2025 equivalent of making students find things using index cards in a library."
I don't think this is a good analogy, it's bad like the analogy with calculators is bad.
Grammarly is now an enemy of education
"Universities risk putting students in impossible positions where they’re simultaneously told they can use a tool and that using certain functions of that tool constitutes cheating."
Seriously, ban it. It's no friend of education
wonkhe.com/wonk-corner/...
looking for some narrative juice this weekend? Most of our games are on a discount right now - with 80 Days for less than a bus fare.
Behold: the first-ever list of news outlets that have banned generative AI in their reporting. As of today, this is literally information that you cannot find on Google.
My goal is to fill the starter pack, so please send over suggestions with supporting evidence!
go.bsky.app/8cn1XfT
21:15 S02E05 Gone March 30 2001
