Dr Sarah @drsezzer.bsky.social · 11hNope, it's me, I have to adjust my expectations... bsky.app/profile/bram... 100
Dr Sarah @drsezzer.bsky.social · 13hEvery time I read something about Anthropic, I'm surprised at how daft they are...the goal was to allow their models to play with webpages, what did they think they'd do?? Other than waste other people's time and resources! Daft, arrogant, naive and/or careless? 110
Reposted by Dr SarahAnil Seth @anilseth.bsky.social · 10/10/20261/2 🎇 Good news! 🎇 The kind people at Behavioral and Brain Science have made the new collection on "Conscious AI and biological naturalism" #OpenAccess for a limited time: target article, all 50 commentaries, & my response. Go to the link and click 'save pdf': doi.org/10.1017/S014.... Do share! 211250
Dr Sarah @drsezzer.bsky.social · 09/10/2026People are assuming this is to clean up the chats for better training data. I suspect it's part of their play to maintain an air of personhood, purely to add an air of mystery/magic/impending doom to their models. Silly, performative, and cynical. Unless they actually believe it!? 110
Reposted by Dr SarahGrady Booch @booch.com · 07/10/2026Ms Hamilton coined the term “software engineering”. 8660209
Dr Sarah @drsezzer.bsky.social · 08/10/2026It's good to see they're sticking to their values of - the rich getting richer, though! (That was sarcasm, she's as daft as she is annoying!) 000
Dr Sarah @drsezzer.bsky.social · 08/10/2026This lady invented the field of software engineering, and saved the Apollo mission...! More than a hint of misogyny from the BBC here! 🤬 BBC News - Margaret Hamilton, whose software helped land Apollo 11 on the Moon, dies at 90 www.bbc.co.uk/news/article...bbc.co.ukMargaret Hamilton, whose software helped land Apollo 11 on the moon, dies at 90Pioneering computer scientist led the development of the software for Nasa's Apollo mission and was awarded the Presidential Medal of Freedom. 000
Dr Sarah @drsezzer.bsky.social · 02/10/2026I also have found her rather close minded and more than a little hostile. Strange way to behave when you're looking to persuade people to you side of the argument. I rated her earlier work, but now avoid her content. There are many others equally competent in this space, and less ivy-towered! 160
Reposted by Dr SarahHypervisible @hypervisible.blacksky.app · 29/09/2026“What if we released malware represented in the form of a cute avatar” is an actual business plan now.engadget.comDots are OpenAI's new personal agents and soon you'll be able to control several of them - EngadgetSure, seems safe. 611183384
Reposted by Dr SarahSam Freedman @samfr.bsky.social · 29/09/2026Was interviewed by a US journalist today who asked if we were all panicking about terrorists at Fairford, which ended with me trying to explain dogging. 551071147
Dr Sarah @drsezzer.bsky.social · 29/09/2026Happy 'starting a new play through' day to all who celebrate! 000
Dr Sarah @drsezzer.bsky.social · 28/09/2026Hah! ChatGPT just swore... "Work out what the bloody thing is doing?" Never seen that before!! (I have sworn at it, but that was in a different session, with memory off!?) 120
Reposted by Dr SarahMartin Rak @martinrak.bsky.social · 04/07/2026A Midsummer Night’s Dream After several attempts, I finally managed to photograph fireflies in the Czech woods. 1124009502
Dr Sarah @drsezzer.bsky.social · 22/09/2026"But their imposed identities don't have all that much influence on task execution beyond shaping their style of text output." They beat us to it, but this lines up with our findings too! 000
Dr Sarah @drsezzer.bsky.social · 17/09/2026And y'all are worried about the consciousness of an algorithm that laughs at jokes! 020
Dr Sarah @drsezzer.bsky.social · 16/09/2026I don't understand how people are using LLMs to deliver research projects autonomously, if they care about accuracy! I've just spent a day or so recreating an experiment (a fairly simple one), but most of the time has been devoted to stopping it spin a n 'impactful' conclusion... despite the data! 010
Reposted by Dr SarahMark Riedl @markriedl.bsky.social · 15/09/2026METR vibe-coded a website and expose an API key that gave an attacker access to $600k worth of compute. But don't worry, METR has been named by Dario Amodei as one of the organizations that is going to audit AI safety at Anthropic. thehackernews.com/2026/09/atta...thehackernews.comAttackers Steal METR API Key and Consume AI Credits Worth About $600,000Attackers stole a METR API key and consumed credits worth about $600,000, while a later campaign failed to access internal data. 0607
Dr Sarah @drsezzer.bsky.social · 15/09/2026This but also, no one talks about scale... The open ai / hugging face attack, was like 700 agents running for weeks, that's a cost noone can sustain unless you own the models. Also to replicate a frontier model needs very large servers to run on... Again, a nuance that's in the hype/fear/noise! 110
Dr Sarah @drsezzer.bsky.social · 15/09/2026Agreed as someone who works in the field... It's so frustrating to hear someone mix up the terms but the worse issue is the constant over anthropomorphising... No wonder it's scary to the vast majority working outside the field! 010
Reposted by Dr SarahThe Tennessee Holler @thetnholler.bsky.social · 13/09/2026“Guys, they want us to go home.” 🔥 🇫🇷 French short film about what would happen to society if all the immigrants listened to angry old racist white men and left (aka the Republican agenda) 15246372045
Dr Sarah @drsezzer.bsky.social · 13/09/2026Urgh. He downloaded GPT and instantly assigned it 'she'. Then she laughed at his joke... (You can't make this shit up!) Wheres a wall I have to go bang my head in it!!! 031
Dr Sarah @drsezzer.bsky.social · 13/09/2026From a human psychology pov this has to be fascinating. Group think at a new level! Hot housed by uni / startup then openai/Anthropic. I would suggest minimal life experience due to minimal time away from thinking about work. And when they try to step away, media attention==instant nerd fame! 031
Dr Sarah @drsezzer.bsky.social · 13/09/2026...against the primary goal. Indeed, the mass production at faster and faster pace of "true/false" statements could destroy fertile ground instead of breathing life into new ideas." Different domain, same loss of cognitive understanding, that hits and will continue to hit agentic software dev! 010
Dr Sarah @drsezzer.bsky.social · 13/09/2026"In recent months, the success of AI in solving major mathematical problems has made headlines even outside mathematical circles. But solving problems is only a tool and proxy for achieving the primary goal of conceptual understanding and insight. Forgetting this in the world of AI may turn the ... 110
Dr Sarah @drsezzer.bsky.social · 13/09/2026The missing piece from all this coverage is the harness. People focus on the LLM, but we need to engineer better harnesses, which reduce inaccurate generations, through context management, and constrained tools. The capability of the whole system is different to that of the LLM! It's not covered. 030
Reposted by Dr SarahMark Riedl @markriedl.bsky.social · 09/09/2026GPT-Astra becomes the first model to be able to beat Zork x.com/ccui9/status... 4143
Reposted by Dr SarahJack Stilgoe @jackstilgoe.bsky.social · 05/09/2026I’ve gone in search of that group of cosplaying men with a perverse idea of national pride who are holding up traffic. I think I’ve found them 916927
Reposted by Dr SarahDaniel Moser @dmoser.bsky.social · 03/09/2026Pollution levels in London after 10 years of advancing sustainable transport. 164115623175
Reposted by Dr SarahTom (he/him) @tompowell01.bsky.social · 04/09/2026if this gets a couple of days of headlines and then disappears, while Arday generated something like the 250-article avalanche, that contrast is going to be pretty glaring 233325
Reposted by Dr SarahSung Kim @sungkim.bsky.social · 03/09/2026Humans are back! Shin Jin-seo, the world's top-ranked Go player, on Tuesday completed a dramatic comeback against the world’s premier artificial intelligence Go engine, KataGo, claiming a historic human victory over AI. www.kedglobal.com/artificial-i...kedglobal.comGo grandmaster Shin defeats AI KataGo in historic human victory - KED GlobalShin Jin-seo, the world's top-ranked Go player, on Tuesday completed a dramatic comeback against the world’s premier artificial intelligence Go engine, K 125450
Reposted by Dr SarahThe Witcher @thewitcher.com · 04/09/2026Back on the Path 🐺 Screenshot by: x.com/Keng_console 123427
Reposted by Dr SarahCount Binface @countbinface.bsky.social · 03/09/2026At what point does a British by-election get awarded to the silver medalist, à la Ben Johnson and Carl Lewis? Asking for a friend of a friend, i.e. me. #Clacton 994986866
Reposted by Dr SarahMerriam-Webster @merriam-webster.com · 31/08/2026macabre = “muh-KAHB” Worcestershire = “WOOSS-ter-sheer” omnipotent = “ahm-NI-puh-tunt” reticent = “RET-uh-sunt” epitome = “ih-PIT-uh-mee” myriad = "MEER-ee-ud" Never make fun of someone if they mispronounce a word. It means they learned it by reading. 795166633669
Reposted by Dr SarahThe Witcher @thewitcher.com · 30/08/2026Witchers, need more The Witcher 3: Wild Hunt – Songs of the Past? Presenting a showcase supercut of our gamescom 2026 demo - 12 minutes of gameplay, monsters, and a peek into the world of Letten in this next expansion! 📷 Watch it on YouTube in 4K: youtu.be/G3kuMWediSQyoutu.beThe Witcher 3: Wild Hunt — Songs of the Past | gamescom 2026 Showcase SupercutYouTube video by The Witcher 217340
Dr Sarah @drsezzer.bsky.social · 25/08/2026Indeed the whole premise of the original post was, 'Hah, when allowed to move in their own way, they're more effective' (if you consider not blowing up/over heating effective). And this misunderstanding is exactly why over anthropomorphising them by running a human race track is misleading. 000
Dr Sarah @drsezzer.bsky.social · 25/08/2026The point is, and I do see what you're saying. Is that they're robots, different weight distribution, different strengths, different joints. They may be able to improve what their bodies are capable of by adopt new styles, but that will not cross over to humans as we are different! 100
Dr Sarah @drsezzer.bsky.social · 25/08/2026Sporting norms for the human body, aren't going to be overturned by a machine. 110
Reposted by Dr Sarahc0nc0rdance @c0nc0rdance.bsky.social · 22/08/2026Finnish nature photographer Lassi Rautiainen followed a female grey wolf (Canis lupus) & a male brown bear (Ursos arctos) for 10 days. They played together, shared meals, napped together. Both were young adults. They appeared to be... friends? (📷: Lassi Rautiainen, www.wildfinland.org) 643339772