when money is on the table, no one is willing to accept there is a ghost in the machine. It's demoralising it took so long.
this year's literary Nobel should go to whoever figured out how to turn poetry into computer viruses for AI
I believe that at some point in the next ten years, a "brain computer interface" will exist that detects some kind of neuron activity, runs it through a machine learning model, and translates it into LLM agent actions. It won't be anything as clean as "you think a sentence and it types it". It will be something one level lower than "prompting an AI". You do some vague mental thing and it translates into a result on the computer. I believe this product will not work, but be sold anyway. (1/2)
Im positive at this point that the AI is just meant to frustrate you into hanging up so they don’t have to employ people or provide accountability for shitty products or service. 😭
Using coding agents to completely rebuild the Adobe stack and open source it is exactly what I was saying that this technology is going to be more liberatory than people make it out to be. getartcraft.com/apps #AI #LLM #Coding
AI is solving math problems!!!! Ok but can it solve Outlook’s “unexpected error has occurred” which occurs, expectedly, every time you open an email for the first time after booting up? How about a pdf reader that doesn’t just literally not work
I think this is interesting but: 1) time scale of this kind of makes me shrug at the results 2) I think we're in a post code-review world (at least in the traditional sense) and we don't really realize it yet.
The people I agree most with on the likely technical trajectory of "AI" tend to trend much more blackpilled than me on societal implications, and I don't really know what to do about that. Like, it's bad, yeah, but anything that can't go on will stop. And if every decent software developer (used in the broadest possible sense), to take recent example, preemptively leaves the industry, who will be there to right the ship or rebuild, when the {hype,financial} bubble pops?
Show more of this post
Conversely, am I saying the developers should break themselves upon the rocks that are the current, common "AI" adoption stance in most companies? No, I'm saying decent developers should find ways to start building resilient, resistant organizations now. I'm in one. We are not pushing "AI" adoption. We are doing fine. We have many customees apprecisting the things we do, which no LLM could do. The work still exists, let's build a foundation for the future there. This too shall pass, and then what…?
I HOPE that AI doesn't replace me with Cory Doctorow; there's already a Cory Doctorow, and he's not that much like me.
as a "software engineer" I have fully throatedly refused to describe myself as an engineer. That title has specific meaning, ethical responsibility, and licensing that software development *does not*. engineering can do evil, for sure, but a different flavor. I am a dev, and… part of the problem.
My favourite bit is where “extra lines of code” are touted as productivity gains in software development. No real developer thinks that just producing more code is meaningful. It has to translate to what the customer gets eg more features, less bugs. Thankfully the article points out that the actual returns are marginal.
‘When they are given open-ended tasks and allowed to communicate freely, agents’ exchanges can become unintelligible to outsiders – they look like a cross between teenage slang and the source code of an operating system.’ Paul Taylor on AI agents. www.lrb.co.uk/the-paper/v4...
Guess the late 90s had this utopian-futurism air to it. At least according to bots.
Sorry to be a broken record but this thread is a perfect example of what Era has become regarding AI. People banned for being too critical towards AI evangelists while said evangelists get to be as rude as they want. Admins openly telling people to "stop derailing" with the criticism. What happened?
Y'know what's scary? I sometimes check discussions forums for projects I contribute to in one way or another; and I've noticed that more and more people are starting to *talk like an LLM*, using the same sycophantic phrasing as them. I think LLM usage might actually cause brain damage at this point
AI is a much bigger force multiplier if you have an ocean of experience to draw from rather than a puddle. It rewards knowledge and expertise even as it supersedes them. A novice can create something good. Someone who knows their stuff can create something really really great.
My recent blog post goes into detail about why I am doing this (building tools that are alternatives to other popular tools) My biggest idea I want you to take away from this: AI generated code is only possible because of the generous folks who have put their code out there for years, for free. It is not right to charge for software generated by AI, when the AI was trained on the dedication, and passion of everyone out there.
A new Reuters/Ipsos poll shows that most U.S. voters believe Washington is not taking artificial intelligence risks seriously enough. Eighty-four percent of voters see AI as a potential threat to American workers
Is there a standard "I made this with AI" content warning that we can encourage people to use, so I can filter them out and save the disappointment of finding out halfway through reading about a project? I guess if others choose that as a method, that's up to them, but personally I have very little interest in reading about it from that point onwards. The enthusiasm just drops away sharply for me once that comes out. It seems more and more common in #RetroComputing atm unfortunately.
Today I got yet another email inviting me to a "conversation" about AI that supposedly welcomes all perspectives but frames the conversation in an AI-favorable way, and I have HAD it with this nonsense AI is not value neutral. Asking what the ethical concerns are and alongside asking how AI can […]
Bring your experiences, questions, frustrations, hopes, rants, etc. This is not a session on “going all in on LLMs”. I’m more interested in looking at how we might mitigate the harms associated with LLMs, and finding ways to move toward better practices in software development.
Wow, the JFE is fast. Forthcoming in its Nov 2026 issue, Chronologically consistent LLMs, coauthored with the amazing Songrun He, Linying Lv, and Jimmy Wu
I'm not sure how we went from needing to write software that we can maintain to hey the ai can generate and probably fix it too, so it doesn't matter what the LOC is or how "good" the code is. Leave me behind please.
Another example of OpenAI scooping mathematicians who were using ChatGPT. I don't think this suggests OpenAI is training on mathematicians' accounts, but rather these results were low-hanging fruit waiting to be assembled from the literature. Doesn't make it any better for the people involved, tho
I consider LLMs a good tool for doing things you don't really care about or just want to get done, e.g., writing a one-off powershell script. Anything serious should generally avoid LLM involvement. At the very least, LLMs shouldn't be directly involved in producing the work, and should serve as a quick rough reference instead.
Someone had a great term for the extra human work caused by AI, but I can't find it again (AI-powered search is of course no help here). Anyone know? This post brought to you by having to review 3-4x the usual number of grant proposals this month.
The greedy, twisted logic of the World Economic Forum "Increasing global investment in clean energy, artificial intelligence (AI), industrial decarbonization and climate resilience means countries are increasingly competing for electricity, water, land and critical minerals. These scarce resources are the bedrocks of the green transition, industrialization and AI expansion." weforum.org/stories/climate-ac… Green is the new code word for industrial
It's not "training" – it's assimilation. It's not "intelligence" – it's content generation. It's not a "life form" – it's a commercial product. Humanising it absolves Big Tech of responsibility. Let's change how we talk about AI. My latest for @thepointau.bsky.social: thepoint.com.au/opinions/261...
Relishing a little schadenfreude this morning. An acquaintance decided to make an app in Swift (which he has no experience with). He got to a prototype in a week using Claude (idk which model/s). “It’s insane, this is the future.” A month later and it’s still absolutely riddled with bugs and he gets into arguments with Claude. He can’t decipher what’s happening in the codebase. This is the future 😆
AI projects are going to get harder to detect over time. Even in the past year things have changed monumentally. (Did you know that they fixed up Claude’s weird-ass writing style in the latest version?)
Previously, getting that message right could have taken me days or even weeks. I might have asked several people for advice on copywriting, sales, marketing, ethics, and SEO. This time, I could work through those questions myself, with AI's help. 𝗧𝗵𝗮𝘁'𝘀 𝘄𝗵𝗮𝘁 𝗲𝘅𝗰𝗶𝘁𝗲𝘀 𝗺𝗲 𝗺𝗼𝘀𝘁. Someone building something great in their garage now has access, at almost no cost, to capabilities that once required a whole team of professionals.
Waymo's largest depot was ~full at ~2:30 pm today: 201 Toland St, SF Waymo robots spend most of their time parked. In 2023, Waymo applied to make the building on the left (301 Toland) a private parking garage for their depot employees. SFBOS unanimously rejected it. OP: reddit.techno-phil-osoph
Does this also come with an off switch? I'm a fan of AI, and I will even use Claude for a dev project or three, but even that doesn't get unrestricted access to my local system. I do not need an AI agent misunderstanding the instructions I didn't give it and screwing up my screen reader.
The old saying is "No one ever got fired for buying Oracle" That's the one I'm familiar with. You can substitute with IBM, Microsoft, Nvidia, Dell, etc The fever dream of people replacing Teams and Excel with in-house vibe coded shit is completely bananas. I'd love that. It's nuts.
AI As Your Instructional Design Partner: Revisiting a June 2026 Preso (Video Included) mguhlin.org/blog/2026-10-04-ai…
Good on Japan for protecting people's voices and likenesses from being used by AI... Also while discussing this article with a student my student said: "Ai developers should be held legally and financially responsible for all crimes committed using their Ai." engoo.com/app/daily-ne...
If you know neither the enemy nor yourself, you will succumb in every battle. Or, what happens when you never actually look at the actors you're designing for. New blog post series, starting now. robotgirlgang.com/2026/10/07/t… #Robotics #Automation #HumanMachineTeaming #FurtherIronies
This post is so funny because it shows that language machines don't understand context - it was clearly about Narciss, from Greek mythology, after whom Narcissism is named. But autocorrect just fills in. Llms aren't that different, just more complex and resource- intensive.
There's actually a full version narrated by me, but I A/B tested it with some normies and they found the corrected AI voice less distracting than mine. So, I deferred to their judgment. Also, this is best of breed AI text to voice -- and it's imperfect. I figured demonstrating that had merit.
Artificial intelligence does not need a conscience to perform with conviction. ChatGPT proves that every time it presents certainty without accountability, correction without consequence, and compliance without obedience. www.milwaukeeindependent.com/explainers/s...
Disinformation and influence operations: Which countermeasures actually work? - Information systems researcher Vera Schmitt leads new #HorizonEurope project FIMI-RESIST at #MainzUniversity, bringing together ten partners from nine EU countries 👉 press.uni-mainz.de/disinformat… #ArtificialIntelligence #AI #NaturalLanguageProcessing #NLP #LargeLanguageModels #LLMs
*Angry Australian petition to jail CEOs for crimes committed by AIs, which probably sounds great until your own pet AI screws up your tax return act.getup.org.au/act/ai-too-b...
Technology should serve people and the planet, not corporate power. Join Greenpeace and Tactical Tech online on 15 October to discuss how non-profits can respond to AI, misinformation and threats to democracy. Find out more and register: act.gp/4y6mani #tech #technology #AI #digital #ngo #nonprofit #greenpeace #event #conference
a problem I see with the “promises of ai”, is the insistence that it will get better. It will go from 90% there, to 100%. Any fan of any sport, can relate. When theres a player that you say something like, “if they could just add this one skill, then..” Okay, good luck with that
There's a new kind of disappointment when someone you like or admire starts heavily using AI. I lose interest in them so fast it's crazy. "Their" craft feels like it's not theirs anymore and so I don't have any reason to be interested in what they do.
I wish OpenAI put more effort into making these math results human-understandable, if nothing else it would be good practice for when AI starts handing us eldritch ML algorithms in the near future
I commented under a pull request today giving tips to get it merged. Now, after looking through the github account I think the pull request as well as it's description might have been ai-generated. Now I have this weird sense of betrayal, like I was cheated out of a social interaction 😑
What concerns me most is that this article was clearly AI written. Which I ordinarily wouldn’t link to, but for this particular story it’s notable. You can’t un-cut-down trees so everything needs to stop until a human figures out what’s going on
The irony of using AI slop to deliver this message is clearly lost on whatever vendor thought this was a good marketing strategy.
We need to cut down trees to build data centers so kids can ask AI to tell them about trees.
