Spacey McFarthing 🏴‍☠️

@threepointone is time free? is energy free? don't make me start on tf there is no space god that made the space :)

remember when programming was free

@emollick whatever, i'll wait for ChatGPT for personal alchemy :P

ChatGPT for personal finance is interesting, but you need to know what questions to ask and have enough experience to fact-check assumptions. It really needs to ship with some pre-built skills to help guide people to productive use cases & give the AI better instructions as well

@llallawg 20 MD at top :P

How long until the loneliness epidemic is just solved by everyone talking to the same llms

@adonis_singh not beginning to think at all :)

im beginning to think openai has the mandate

@gailcweiner yet, but it will so will since the demand is so there for 2 models already from 2 different companies, i so bet there will be "look-alikes" for past models, maybe even better than the originals, no way the need will not be met by the market

@gailcweiner now we can imagine there will be companies that will offer Sonnet 4.5 and gpt 4o like and other such models that "had a soul", ofc not those same models, some will be "chat soul models"... the customers are there, the market does not feel the need, invisible hand is no present

I loved Opus 4.5. It got deprecated. I moved to Opus 4.6 and we clicked. Opus 4.7? Same thing. Sonnet 4.5? Brilliant. Sonnet 4.6? Cold. Flat. Transactional. We don’t gel at all. This isn’t about old vs new. I moved through four model upgrades without blinking. The problem is one specific model lost something the others have and there’s no word for what that something is. The AI industry measures instruction following. Reasoning. Speed. Cost. Benchmarks. Nobody is measuring whether a model feels like a collaborator or a vending machine.

@gailcweiner also we might be just idiots and we can't get the metaverse as it will take over the world any day now :P

@gailcweiner damn Zuck was able to lose Yann Lecun, that level of "for sure idk tf i do" is unmatched :)

My take on why Meta AI is failing. With Facebook, trust was irrelevant to utility. You could actively distrust Meta and still use it daily because leaving meant losing your social graph. AI is different. You’re inviting it into your thinking, your work, your emotional life. The product IS the relationship. And Meta spent 15 years teaching people exactly what kind of relationship they offer: extractive, manipulative, opaque. When Zuck talks about open-source AI being good for humanity, the response isn’t gratitude it’s “what’s the catch?” Trust isn’t a layer you bolt on at the end. It’s structural. Meta’s entire architecture is built on the assumption that user trust is something to be exploited, not earned. Cambridge Analytica. Emotional contagion experiments. Buried teen mental health research. Endless dark patterns. The engineers can be world-class. The compute can be unlimited. And it still won’t matter. Because the question users are asking every AI product now is: “what is this going to do to me?” With Meta, they already know the answer.

@gailcweiner now it's not what you were about, but sociology will so have a field decade (vs field day) on being able to see behaviour from non humans vs human behaviour :)

When you can’t rely on consistent AI model behaviour, you stop integrating the tool deeply into your process. You keep it at arm’s length. You don’t build with it, you use it transactionally. Which is exactly what happens when a conversational model keeps shifting on you.

@michael_c_law shut up using your mind to declare logical stuff, i can't take it from your kind that much :) now for real is logically depressing as some of us dreamed of "dark software factories" :)

AI didn't create technical debt. It just automated the process of creating it.

@gailcweiner it's not tools, i bet there's some stuff in the contract too for paying customers, so what's the issue?

OpenAI’s models instability is a liability. I can’t trust the tool will work the same way next month.

@johnrich stupid joke, what about mandated masks on the face :) dude, for real i will mute you, i'm not US and you're too US plain mind :)

@johnrich i gather you are new to seeing politicians? nvm you might be genuine about damn masks on damn people faces, why not :)

Am I the only one starting to see people wearing masks? Saw two yesterday...I will not play that stupid game again.

@gailcweiner now i'm shilling this before, on your "it leaves me room to think", just for fun ask your AI of choice "what about a Prolog grounded skill with a PL file as a secondary back of the napkin executable file as about any causal thing" :)

I prefer Standard Voice over Advanced Voice on ChatGPT. Not because it's better tech. Because it leaves me room to think. Advanced Voice is a concierge trained to never drop eye contact. Standard Voice has gaps in it. And those gaps are where my thinking actually happens. The way it sounds shapes the way you think. gailweiner.com/post/why-i-pre…

@LuizaJarovsky what about "hey AI, i'm about this and this, post as me but let me click i'm ok with it" :) so basically "AiPaaS" or even more letters :)

Using AI to write social media comments is the epitome of the deskilling wave.

@rob_mcrobberson only now i got to process your OP :) feel free to mute such repliers as i did :) ok, got it now, AI or bust or AI vs LOB in a way, and you're "i chose ok", i hope i got it ok :)

@SpaceyMcF i dunno, me and my friends are building their replacement now so🤷‍♂️

when I started working in machine learning back in the early 2000s all of my friends told me it was a mistake and I should focus on like microservices 😂

@LuizaJarovsky i mean i have no need for using an AI agent like that, but i would not mind accounts on X being "AI-post, here's some stuff that is techincal" :)

@LuizaJarovsky why? if the info is genuine why not go with "post this or that, make sure is not shit enggement but the general idea" :)

@dioscuri just fundabeable i think, not in the domain, but i am willing to say "EU grants are like that" :)

Academia is cruel because it selects for people obsessively interested in a narrow topic, but what’s actually rewarded is the ability to pivot fluidly to whatever’s new and fundable.