benwaugh.bsky.social @benwaugh.bsky.social · 05/10/2026It's been a while. I don't know if I've even still got whichever book it was but I'll have a look. 160
benwaugh.bsky.social @benwaugh.bsky.social · 05/10/2026He also referred to video recorders watching TV so you didn't have to. And one of his characters got rich by creating software that came up with justifications for your decisions instead of helping you decide what to do, which is pretty much what LLMs are for. 21234
Reposted by @benwaugh.bsky.socialJohn Stott @jpsastro.bsky.social · 24/09/2026Had a UKRI funded grant proposal rejected at the "AI triage stage". "This AI-based triage was used only to identify approximately 50% of the strongest proposals to take forward to full human review." 🧪 #academicsky 40435241
Reposted by @benwaugh.bsky.socialBy Cory Doctorow (GPG 0xBF3D9110957E5F4C) @doctorow.pluralistic.net · 11/03/2026"We used an AI to do this" is increasingly a way of saying, "We didn't want to do this in the first place and we don't care if it's done well." 21/ 528682
benwaugh.bsky.social @benwaugh.bsky.social · 26/09/2026FFS. Why would I spend my time reading something the purported author didn't even take the time to write? 010
Reposted by @benwaugh.bsky.socialJo Wolff @jowolff.bsky.social · 26/09/2026Think I’ve found those famous AI guard rails. 56214
Reposted by @benwaugh.bsky.socialPavel @spavel.bsky.social · 20/09/2026Software developers think "if AI automates coding, I will be able to spend less time on producing widgets, and more time on strategy and architecture." Bosses think "if AI automates coding, I can get my devs to produce 100x more widgets." 7708150
Reposted by @benwaugh.bsky.socialBruno Dias @brunodias.dev · 21/09/2026In a lot of these spreadsheet-job scenarios is AI providing a sort of (inefficient, unreliable, ethically unacceptable) swiss army knife that papers over poor tooling and generally shitty labor conditions. If a task is empty drudgery why isn't it amenable to normal automation? 613315
benwaugh.bsky.social @benwaugh.bsky.social · 10/09/2026Some good (and well expressed) points here from Sarah Demers about the loss of attribution and of human conversations and learning that can come from using LLMs, while acknowledging the ways scientists are finding them useful. 011
Reposted by @benwaugh.bsky.sociale.w. niedermeyer @niedermeyer.online · 21/08/2026Threatening people with inevitability doesn't work anymore. Overemphasizing the benefits and minimizing the risks doesn't work anymore. Dismissing peoples concerns because they aren't deeply rooted in the technical details doesn't work anymore. The days of steamrolling hesitancy are gone. 742560
Reposted by @benwaugh.bsky.socialChad Loder @chadloder.dev · 24/06/2026Oh now Thiel is talking about the "reproducibility crisis" in science, amazing Wait these guys are all myna birds, aren't they. They hear a few words said by someone else and they just grab them and repeat them without understanding. Now I see why they love LLMs so much. 302364228
benwaugh.bsky.social @benwaugh.bsky.social · 10/06/2026I can't see anything about coal in the linked articles. What am I missing? 110
benwaugh.bsky.social @benwaugh.bsky.social · 29/05/2026A colleague in P&A worked with some students recently to create a proof of concept for something like this. Can put you in touch when back at work on Monday. 010
benwaugh.bsky.social @benwaugh.bsky.social · 21/05/2026Er, did you mean to reply to a different post? 000
benwaugh.bsky.social @benwaugh.bsky.social · 21/05/2026Generating plausible statements has always been the bottleneck in scientific progress, after all. 020
benwaugh.bsky.social @benwaugh.bsky.social · 21/05/2026"Teams of AI agents boost speed of research." in that they can "arrive at plausible hypotheses". Wow. www.nature.com/articles/d41...nature.comTeams of AI agents boost speed of researchSystems can generate hypotheses, interpret data and suggest ways to develop medicines. 130
benwaugh.bsky.social @benwaugh.bsky.social · 03/07/2025I wish the book I'm reading now (about higher education, not programming) had exactly this. It's not a whodunit. Give me the spoilers! 010
