My husband and I (both working in software - me on the product management side, him on the engineer side) were just talking about how AI is able to do some of the fun parts of our jobs but not the drudgery. It is backwards to what you'd want
AI · Takes · page 4
Further down today’s ranking of AI posts that are opinions and questions: page 4 of 4. Posts the model is less sure belong here come last.
If AIs were really there to better our lives, empower us and whatnot, every single model would be open source. But in reality they are little black boxes whose sole purpose is to make money. Can't wait for the day when Claude and co have nothing to train on than the slop they produced themselves.
A lot of the attraction of AI is that it can automatically bypass the service isolation that had been the norm for computing in the last 20 years Nothing communicates to anything else, except through a human - or an ai agent
I need a badge I can carry around that I can present to boosters who assume my skepticism is based on ignorance. “yes, I know what RLHF is. yes, I know what GANs are. yes, I know what RAG is. yes, I know that recent models have made significant gains on various benchmarks. please listen to what I am saying and stop arguing with a guy you made up”
I'm reading @joemoransblog.bsky.social's book on sentences, and his discussion of Socrates' on sentences has unlocked for me something about the fundamental problem with GenAI. It's is very good a mimicking human sentences. But sentences are not what makes meaning, human writers & readers do...
Snarky observation... I see a lot of mathematicians rushing to try to use AI to create some form of human readable document out of the Lean code Levant had Claude produce. A few have created documents in the human scale of, i.e., 100s of pages not 250k. Two things seem plausible:
A lot of discussions about AI taking jobs eventually wind up at: "I'd like to see AI do [job requiring physical labor or presence]." And...ok. Sure. But we can't all be plumbers, & this position cedes whole constellations of creative & thinking work to the machine. I'm not interested in doing that.
This x1000. It is such a rookie mistake for journalists to equate “AI bubble” with “AI is useless.” If it were useless, the bubble wouldn’t inflate this much. Many AI companies are attracting investments that are predicated on the promise of successfully driving each other out of business…
GenAI handwriting recognition is often in plausible but wrong even in English, French, Spanish, and German. The best way to use it is to offer a human transcriber ne line at a time for them to review for accuracy (even with Gemini).
One of the worse aspects of agentic tools like Claude is that they are designed to strive to take away as much agency as possible. "Automate this? It will be faster if you let me to that? Want me to automatically fix this? Should I always take care of that for you?"
It's totally fine to say "A.I. is bad because of [reason]." It is not fine to say "A.I. can't actually do anything at all. The A.I. proof of the Riemann Hypothesis isn't a real proof and wouldn't count even if it was." And I'm seeing some of the latter on here, and from people I respect a lot.
AI’s ethics can be superior to ours precisely because it lacks affective consciousness. Affective neuroscience supports the view that feeling precedes reflection. We can know a rule perfectly and still feel entitled to an exception. What does the machine stand to gain?
So this #radiostation on here…. I called them out yesterday for using AI art for promotions. I see this shit every day and get sick of it. A music radio station using #generativeai for promotional material is gross to me. All I said was it was a big turn off to me. That is it. Then today they started arguing with me and another at least. Lol. Told me i am wrong. “People love the AI photos for 3 years now.” Compared our follower counts. Called me a “so called musician” and saying i was a moron, riding a “bandwagon” because i was turned off by GenAI. Had a whole thread going before they deleted it. Good riddance. For the record. (Puns) I just want to state, This was a reaction to displeasure with #GenAI in the ad. I never listened to the station once. They only show up in my feed because of the #music tag. The response is telling to me. What a reaction. They went all, “my rights!”, over a fake photo… People. What is happening?
It’s entirely possible that an AI bot solved the Reimann hypothesis but if the proof can’t be explained in words to another mathematician then it’s not really a proof at all
I’ve said this before but the fixation on “wokeness gone wild” anecdotes on elite campuses is interesting when stuff Iike this is going on in the second largest collection of public university systems in the country www.theatlantic.com/ideas/2026/1...
I'm beginning to think that AI is causing a split in intellectual society more profound than CP Snow's two cultures.
all those ads all those bots all those trackers all those LLMs exist somewhere on a hard drive which means they can be destroyed
The obvious question is whether there could be leakage from this or something like this into whatever process Anthropic’s model used, and I don’t think even Anthropic itself could guarantee the answer is “no”. If you put something anywhere on the web, you’re in the training set
I'm seeing lots of talk (& adding to it!) of how AI-boosters don't get that art, writing, music, etc. aren't about the *products* of those activities, but about the experience of the activity. That the meaning comes from the effort of creation. But I wonder how many people that's really true for.
Content warningslop image
the "retro TV repairs NZ" group I'm in on facebook (ugh) is starting to get those AI generated slop engagement posts. I usually just roll my eyes, block the OP and move on but this one made me froth. THAT"S NOT WHAT A BETAMAX TAPE LOOKS LIKE!!!!!!!!!!!!
1 photo hidden: marked sensitive
This is so true, and you would think with the rampant pop culture knowledge that the genie will take your wish and twist it if you aren’t explicit about your wishes, that people would understand better how important prompts are. Garbage in garbage out etc
Might as well have AI invent a fake university and print you a diploma from that university. Save everyone the trouble!
I can't help thinking that this sounds like "Michael Phelps has been ranked best prepared to swim through a tsunami" but I guess we'll take it! cmu.edu/news/stories/archives/…
I have been saying this to our privacy lawyer, our IT, our admin for over 2 years. They are bankrupting universities and driving up tuitions by signing these contracts. I sent them a discussion between ex-Microsoft salesmen who compared contracts to selling crack to begin the addiction.
What I believe all non-AI-obsessed people want (including me) is PODS: Plain Old Deterministic Search. I want to enter queries, be shown things that are relevant to that query, and then make a decision about how much I trust it on my own. I am the "inference" machine, whatever that even means.
One might think that all these AI math proofs are a PR stunt to prove the power of an AI model leading up to an IPO and to boost stock prices thereafter ... And I thought that too when 722 papers dropped in Oct with "OpenAI" as the author. That kind of corporate attention will pass with time.
Its also not a novel realisation - ai™️ bubbles bursting, the damage done. It was all visible, and avoidable. People just chose not to hear.
Once again thinking about the @wired.com article where they could not find normal people with stable lives to participate in this experience and everyone involved admitted it wasn’t tenable and/or had a full breakdown.
I don't think AI conditions for how we learn writing will change how people see the teaching & learning of writing because no one outside a writing program cares to understand those conditions. We know how people learn to write but people still see the teaching & learning of writing as remedial.
The day is fast approaching when AI becomes so complex that it becomes impossible to comprehend by humans — this is the end. #AI
It's really funny hearing how things that people have been doing for decades, sometimes centuries, and with some degree of routine, until like three years ago now somehow absolutely require the use of generative "AI" to be possible at all.
I love when people say "my AI", "my ChatGPT". Like, dog, its the exact same as everyone else's. You just have a few tokens in its context
if you purposefully deceive me about using genAI in a contribution I will not hesitate to rip it out and deny you the utility of it github.com/amaranth-lang/amara…
The media is starting to notice #AI is literally ruining the internet. Even just at a security level. Kiki misses the money, but not her job in Tech. You guys can fix it. finance.yahoo.com/technology/a… #tech #fuckai #vtuber
AI targeting system arms dealer, hired Tresor for weapons fair? Dystopian much? theguardian.com/world/2026/nov…
There's a long list of reasons for not liking current generative machine learning applications. For me, a core one is that this tech has nothing to do with intelligence. It's just brute-forced #statistics. Can you stupid fucks stop applying my as yet unattained #scifi tech term incorrectly?! 🙄 #ai #llm #aibubble
Weather Presenter is a job that must be threatened by Artificial Intelligence.
IT'S FINALLY. FUCKING HAPPENING. WE WERE FUCKING RIIGHT. IF YOU WERE PUSHING IT, ESPECIALLY ON YOUR EMPLOYEES, ENJOY YOUR NEW HOME IN THE FECULENT POOL YOU CREATED. LET THE FINAL IMAGE TO HIT YOUR LENSES BEFORE YOUR SYSTEM CEASES IT'S WORTHLESS, PARASITIC HUMMING BE THE PROLES YOU NOW LIVE WITH, THE ONES YOU FORCED TO LIVE THERE PERMANENTLY. BURN IT ALL ❤️🔥❤️🔥❤️🔥🔥🔥🔥🔥🔥🔥🔥🔥🔥❤️🔥❤️🔥❤️🔥 abc.net.au/news/2026-11-21/ant…
theconversation.com/catching-a… What a surprise, more evidence that relying on "GenAI" is actually making people less productive... Maybe we can wrap this around a brick and use it to beat in the skulls of those idiot "C-suite" managers who continue to push "AI" on their workers in the misguided belief that it makes them work better? #fuckAI #AI
Especially when your points are that 1. AI can be trusted and 2. you are an expert at using it, not just some beginner randomly copying and pasting to and from it.
Nice choice of date: 11 November, or Armistice Day. theguardian.com/world/2026/nov…
More in AI
The model is less sure these belong here, so they sit outside the ranking.
Do you remember when AI wasn't designing our cutlery and 3D printing crappy forks? Simpler tines.
It's basically Total Recall's Mars.
Amazon's tech, like the #spectacle itself is a state science.... whether it's open or not it's a royal science and technology ... using AI in all ir any of its forms is evidence of the users desire to be oppressed
#rockstargames straigt up lying about not using gen ai for #gta6 but using gen ai assets in the #furry base. The one place it's definitely going to be found, because we love our artists....