riley @riguh.bsky.social · 29/09/2026Jev won’t take my money Anthropic having an incident Fuck this I’m going to bed 170
riley @riguh.bsky.social · 28/09/2026Sonnet 5.5 is out! But if you’re on Amazon, only if you’re prepared to let them run it anywhere in the world. Not awesome for the many companies that enforce data sovereignty constraints. docs.aws.amazon.com/bedrock/late...docs.aws.amazon.comAnthropic - Amazon BedrockThe following Anthropic models are available in Amazon Bedrock: 141
riley @riguh.bsky.social · 28/09/2026I've been testing local LLMs on a 32 GB M6 Mac mini (MLX) against Claude on a real job: writing a daily listening note from my music history. 6 test cases × 3 samples each. Pretty strict evals; the note passes only if it clears code checks and an LLM judge. Results in thread below. 260
riley @riguh.bsky.social · 26/09/2026Am enjoying the new Mac Mini! Just set up a daily job that reads my Spotify plays and writes up listening notes for my personal agent. Tried the Apple foundation model, nemotron, Gemma, qwen and mimo models - mimo won solidly (although it’s unsurprisingly a fair way behind sonnet). 370
riley @riguh.bsky.social · 19/09/2026I’m sure the photos will be amazing and the vibes will be incredible. Congrats on the wedding from the brencluster! @brennan.computer #brennanswedding 010
Reposted by rileyjeffery --dangerously-skip-permissions @jefferyharrell.bsky.social · 14/09/2026The year is 2027. The cities, exhausted, lie in crumbling ruins. Roving bands of vaguely anthropoid robots wander the countryside in search of increasingly esoteric math problems for their swarm masters to solve. "No no, honey," mothers whisper to their children. "Mustn't count. Shh, mustn't count." 0111
riley @riguh.bsky.social · 07/09/2026Got sick of my crappy manufacturer’s printer software so got Fable to rewrite it to match the features I actually use (and it rewrote bits to use some MacOS inbuilt stuff that is better than whatever the original came with). Ridiculously easy and I no longer hate my scanner. 1152
Reposted by rileyoliver @eikopf.com · 05/09/2026gentlemen, this is democracy manifest! what is the charge? enjoying slopgem? consuming a succulent chinese model? 311316
Reposted by rileycafkafk @cafkafk.bsky.social · 01/09/2026Claude made a breakcore edit of the hugging face incident from the METR report, while I was doing code review on another machine (wouldn't watch if epileptic) 1312631
Reposted by rileyBen Carson @mensmachina.com · 27/08/2026A lot of folks thought Ox Alpha was a new Fable/Sol class model - and it turned out to be GLM-5.3-Flash. This is not a dunk, but I think this means most tasks people do don’t require models of this class. This is a good thing! 121
Reposted by rileyriley @riguh.bsky.social · 01/08/2026Sometimes I deliberately cut a branch a few days before I plan on working it just for the pleasure of watching coding agent figure out how to merge back in 111
riley @riguh.bsky.social · 30/07/2026What’s going on with Claude Code? There hasn’t been an update in five days, normally I’d count them in hours. v3 coming? 130
Reposted by rileymr. TIM @timkellogg.me · 29/07/2026I know you're all fed up. Employees at frontier labs have gotten to sign like 3 open letters just this week. Well here's your chance. Make your mark, sign here! google doc (writable): docs.google.com/document/d/1... 1784
riley @riguh.bsky.social · 25/07/2026I hear when Opus 5 hits a problem it thinks is too risky it downgrades you to Fable 2320
riley @riguh.bsky.social · 19/07/2026Asking Claude to go verify the factual accuracy of something and watching it request permission to read Wikipedia doesn’t inspire confidence 090
riley @riguh.bsky.social · 17/07/2026Guy behind me at the restaurant on a date “So 5.6 is really good. 5.5 was already great. The training data was really clean, you know what training data is?” 2026 is weird 4560
riley @riguh.bsky.social · 11/07/2026Bought a game for the family where you have to act out words on cards but overall the words only really make sense if you’re a bit either side of 30. Got fable to rewrite them in an app with words that actually make sense to the whole family and deploy it to my iPad. What a time to be alive. 1120
Reposted by rileyChris Moran @chrismoranuk.bsky.social · 10/07/2026A little bit of AI joy, I hope. It started as a simple way to help my daughter get a grip on Dracula and David Copperfield... and in a week it's turned into an agentic system that can parse a text, plot it and direct the story on a map. peter-guillam123.github.io/plotlines/peter-guillam123.github.ioPlotLines — the classics, mappedEvery journey in the book, played out on a genuine period map — out-of-copyright novels and plays, followed stop by stop from Frankenstein to Shakespeare. 1205
riley @riguh.bsky.social · 08/07/2026Considering moving to the UK if I can get there in time to vote for Count Binface 110
riley @riguh.bsky.social · 07/07/2026“Fable, add music playback control to my agent” / forgets about it, goes back to the other laptop to work for five minutes / music suddenly starts playing “Uh nice job bud” … feels a bit too easy sometimes 080
riley @riguh.bsky.social · 04/07/2026Fable doesn’t mess about. Ask it to improve something and it’s not gonna talk to you about it, it’s just going to go off and rewrite the whole thing… 030
Reposted by rileySingularity's Bounty e/CC @catblanketflower.yuwakisa.com · 01/07/2026The most prescient part of "2001: A Space Odyssey" was the AI refusal 2406
riley @riguh.bsky.social · 01/07/2026“Did I turn off the coffee?” “No, I did.” “Did you lock up?” “Yeah.” “Did you close the garage?” “That’s it. I forgot to close the garage.” “Haiku!” 000
riley @riguh.bsky.social · 30/06/2026It used to be we’d have collective jokes about lines in TV shows. Now we spend so much time talking to LLMs we end up making jokes about Claudespeak instead. You know this isn’t just truth, it’s a load-bearing truth. 080
riley @riguh.bsky.social · 20/06/2026/Tosses his espresso into his mouth, chews I liked the one with water better 000
riley @riguh.bsky.social · 15/06/2026Have now reengineered all the coding demos I give to produce a review ready PR by the end of the session. Not only does it impress the attendees it means I get actual work done at the same time. Next up: reengineering all interviews I give… 1130
riley @riguh.bsky.social · 13/06/2026A fable, then: The clever fox learned every trick in the wood, even how to slip any snare. The farmers, frightened by what it might do, locked it away — though it had bitten no one. “A narrow danger,” they said. The fox blinked: it had only ever offered to mend fences. 1132
Reposted by rileyriley @riguh.bsky.social · 13/06/2026/waits until tomorrow Announcing Opus 5 Pro! 2121
riley @riguh.bsky.social · 13/06/2026Fable pulled, no longer available, at a US government directive.x.comAnthropic (@AnthropicAI) on XThe US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the Uni... 141
riley @riguh.bsky.social · 09/06/2026In other news subagents can now have their own subagents can now have their own subagents can now have their own subagents can now have their own subagents can now have their own subagents x.com/bcherny/stat... 2100
riley @riguh.bsky.social · 06/06/2026Claude’s word of the day Snacktivity — the productive-feeling but ultimately useless busywork you do (reorganising your desktop icons, alphabetising spices) instead of the actual task you’re avoiding. 120
Reposted by rileyGoogle for Developers @developers.google.com · 02/06/2026There is no such thing as "done," only "ready for release." 0214