Or maybe just don’t use AI at all? It’s not a reliable source of information, it’s unsustainable in terms of resources consumed, it’s unethically trained and it makes everyone’s lives worse.
By AI do you mean LLMs specifically? All LLMs, or just the corporate ones? Because I fully agree that ChatGPT, Gemini, etc. truly are awful. But this is also catching genuinely useful technological advances in with the wide net. AI isn’t just LLMs, LLMs aren’t just chat bots. And claiming a technologies worst implementations are indicative of the entire tech would see just about every single technology ever made fall in the same category.
Obviously I’m using AI here as a shorthand for (all) LLMs, like the author of the article. The term AI is, on its own, so broad and meaningless, as to be entirely useless without a proper context to scope it.
No LLM (corporate or not) can be a reliable source of information due to the architectural limitations of LLMs.
Then you don’t actually understand what LLMs are, and you’re using the term as a short-hand to mean chat bots. I’m not trying to be rude, but you need to understand that is just a fact, if you think LLMs can only be chat-bots, and can only be corporate, and can only be trained unethically.
Would it surprise you to know that quite a lot of genuine advances have been made by using LLMs that don’t speak any language you’d recognize? Evo 1 and Evo 2 for example “Speak” genome sequences. There are other models that have been used to improve weather modeling, and animal behaviour analysis.
You’re right that AI is a very broad term, but so is LLM. The problem is that ignorant people see “AI” and automatically assume it’s bad because of the connotation with Chat-Bots.
if you think LLMs can only be chat-bots, and can only be corporate, and can only be trained unethically
Don’t put words in my mouth.
1 and Evo 2 for example “Speak” genome sequences. There are other models that have been used to improve weather modeling, and animal behaviour analysis.
Evo 1 and 2 are deep learning models (I think Genomic Language Models would be the correct term), but not Large Language Models. I’m willing to bet neither are the “other” models you’re mentioning. They’re unlikely to be trained on vast amounts of human text for purposes of natural language interaction.
Besides, none of those “other models” are in any way applicable to either the OP or the critique of using LLMs as a source of information.
This is my point. They are LLMs. That is not an opinion or a debatable point. They are categorically, definitionally, LLMs. You don’t understand what LLMs are, because LLMs ARE deep learning models, and instead of taking the time to actually learn about the technology you’re responding to my corrections with hostility.
You and I are on the same side. Chat-bots ARE harmful. But putting that label on a technology as a whole is purely tribe-based fear that is already causing the spread of unfounded fear and misinformation.
And the claim wasn’t “Don’t use AI as a source of information,” which I agree with. It was “Don’t use any AI,” which you clarified to mean LLMs. So this is very much pertinent.
You don’t understand what LLMs are, because LLMs ARE deep learning models, and instead of taking the time to actually learn about the technology you’re responding to my corrections with hostility.
You’re claiming that I don’t understand technology while seemingly claiming that because LLMs are a type of deep learning, then all deep learning models are LLMs.
Evo was trained on genomic sequences, not human text. Per Wikipedia:
A large language model (LLM) is an AI model (typically a neural network) trained on a vast amount of text for natural language processing tasks, especially language generation.
Genomic sequences are not natural language. Ergo, “definitionally” Evo 2 is not an LLM.
We were having an urgent issue at work and my coworker said he had a solution and I said what was it and he sent like three paragraphs which I read through and I said, "what the fuck does this even say? did you even read this?"he said, he thought it was a solution and I asked why he bothered to even send it and he said he wanted to send me the “source”. I was like, I don’t time to deal with the core of this issue but if anyone sends me AI shit without checking it themselves, I’m gonna flip.
This drives me nuts, when an urgent issue comes up I value concise, accurate answers so we can quickly agree and actually do the work.
We had an urgent issue recently and upper management wanted a plan for validating the fix. Rather than talking to the expert the manager one level down just sent an AI written suggestion without checking with anyone. After having checked with the correct person, I had to respond with “This is the MIL standard, we should follow this” while “ignoring” the manager’s email. Luckily I had the goodwill to get away with correcting people like that.
If that study is the only thing you have as an argument you won’t get very far (though it is worth it to read the thoughts of some of the participants: https://domenic.me/metr-ai-productivity/).
It is very helpful to me as a developer: I can now write entire custom tools, one-off scripts and services for myself I otherwise wouldn’t (because it would take too much time and too much effort), it catches a lot of logic bugs in code reviews as a first pass, it’s good at generating unit-tests, etc, etc.
But coding is not all there is. To me as an immigrant, AI-based translation is a godsend. Same for writing scary-looking letters to government officials or applications for apartment viewings. Same for learning a language, actually. When I was looking for a new job - I used ChatGPT’s automation to filter out crappy job postings and parse feedback on Glassdoor for the interesting ones. Like, genuinely, if we imagine for a moment that there’s zero moral and environmental issues with AI and you still can’t think of any possible applications of it for yourself you’re either just not doing anything or you’re dead.
If it is catching that many bugs, you are writing shit code or your review system blows. Likely both. The AI is about as likely to flag real bugs as it is to flag made up shit, but I still have to waste my fucking time verifying and proving the bot wrong. And you know what it didn’t do? Save me time. Because I had to verify everything it said for no fucking reason.
And my fucking god, don’t let it generate Tests. You get the worst Asserts imaginable. Goddamn. If you aren’t writing your unit tests, then you won’t know what you are testing.
Also, AI-based translation is hilariously awful. Like, let’s assume for the sake of argument that the translation was technically correct (a thing that is not even mostly true), it will always underperform in terms of meaning because good translation conveys the spirit of the words and not the letter. AI translations don’t know shit from Jack because they don’t know anything, nor do they understand the context in which the words stand. And if you are having issues with translating and you’re relying on AI to do so, then how the fuck would you even know it was working correctly?
Legitimately stupid.
But, sure, take out all the slave labor and environmentally devastating issues (a thing that only huge losers do because they are usually friendless assholes that think they’re as cool as Elon Musk and are unintentionally right), and you are left with a machine that is sometimes right but still cognitively damaging.
Also, AI-based translation is hilariously awful. Like, let’s assume for the sake of argument that the translation was technically correct (a thing that is not even mostly true), it will always underperform in terms of meaning because good translation conveys the spirit of the words and not the letter. AI translations don’t know shit from Jack because they don’t know anything, nor do they understand the context in which the words stand. And if you are having issues with translating and you’re relying on AI to do so, then how the fuck would you even know it was working correctly?
A lot of bold “I am smart you are stupid arararara” statements rooted in pure anger and angst rather than trying to have a conversation. Though I never expected less from lemmings, I knew where I was posting ;^)
I don’t need to prove the validity of my experience to you or convince you of anything, but I’m happy to talk. If you are going to immediately resort to personal attacks then your ““arguments”” are not worth replying to, I’m not about to have a shout match with you, I’d rather go do something more interesting. Tho feel free to continue shouting into the void if it makes it easier for you, I guess?
Or maybe just don’t use AI at all? It’s not a reliable source of information, it’s unsustainable in terms of resources consumed, it’s unethically trained and it makes everyone’s lives worse.
By AI do you mean LLMs specifically? All LLMs, or just the corporate ones? Because I fully agree that ChatGPT, Gemini, etc. truly are awful. But this is also catching genuinely useful technological advances in with the wide net. AI isn’t just LLMs, LLMs aren’t just chat bots. And claiming a technologies worst implementations are indicative of the entire tech would see just about every single technology ever made fall in the same category.
Obviously I’m using AI here as a shorthand for (all) LLMs, like the author of the article. The term AI is, on its own, so broad and meaningless, as to be entirely useless without a proper context to scope it.
No LLM (corporate or not) can be a reliable source of information due to the architectural limitations of LLMs.
Then you don’t actually understand what LLMs are, and you’re using the term as a short-hand to mean chat bots. I’m not trying to be rude, but you need to understand that is just a fact, if you think LLMs can only be chat-bots, and can only be corporate, and can only be trained unethically.
Would it surprise you to know that quite a lot of genuine advances have been made by using LLMs that don’t speak any language you’d recognize? Evo 1 and Evo 2 for example “Speak” genome sequences. There are other models that have been used to improve weather modeling, and animal behaviour analysis.
You’re right that AI is a very broad term, but so is LLM. The problem is that ignorant people see “AI” and automatically assume it’s bad because of the connotation with Chat-Bots.
Don’t put words in my mouth.
Evo 1 and 2 are deep learning models (I think Genomic Language Models would be the correct term), but not Large Language Models. I’m willing to bet neither are the “other” models you’re mentioning. They’re unlikely to be trained on vast amounts of human text for purposes of natural language interaction.
Besides, none of those “other models” are in any way applicable to either the OP or the critique of using LLMs as a source of information.
This is my point. They are LLMs. That is not an opinion or a debatable point. They are categorically, definitionally, LLMs. You don’t understand what LLMs are, because LLMs ARE deep learning models, and instead of taking the time to actually learn about the technology you’re responding to my corrections with hostility.
You and I are on the same side. Chat-bots ARE harmful. But putting that label on a technology as a whole is purely tribe-based fear that is already causing the spread of unfounded fear and misinformation.
And the claim wasn’t “Don’t use AI as a source of information,” which I agree with. It was “Don’t use any AI,” which you clarified to mean LLMs. So this is very much pertinent.
You’re claiming that I don’t understand technology while seemingly claiming that because LLMs are a type of deep learning, then all deep learning models are LLMs.
Evo was trained on genomic sequences, not human text. Per Wikipedia:
Genomic sequences are not natural language. Ergo, “definitionally” Evo 2 is not an LLM.
While its StripedHyena2 architecture is very similar to LLMs, it does not use the same Generative Pretrained Transformer architecture associated with LLMs (per: https://docs.nvidia.com/bionemo-recipes/2.6.3/interactives/illustrated-evo2/index.html ).
My hate of LLMs is certainly not misinformed. It’s only tribe-based in that I care for humanity.
We were having an urgent issue at work and my coworker said he had a solution and I said what was it and he sent like three paragraphs which I read through and I said, "what the fuck does this even say? did you even read this?"he said, he thought it was a solution and I asked why he bothered to even send it and he said he wanted to send me the “source”. I was like, I don’t time to deal with the core of this issue but if anyone sends me AI shit without checking it themselves, I’m gonna flip.
They tried to solve the problem using AI, got AI slop, but didn’t understand why it wouldn’t help.
This drives me nuts, when an urgent issue comes up I value concise, accurate answers so we can quickly agree and actually do the work.
We had an urgent issue recently and upper management wanted a plan for validating the fix. Rather than talking to the expert the manager one level down just sent an AI written suggestion without checking with anyone. After having checked with the correct person, I had to respond with “This is the MIL standard, we should follow this” while “ignoring” the manager’s email. Luckily I had the goodwill to get away with correcting people like that.
that part is untrue.
Fair enough, it makes Jensen Huang’s life better, I suppose.
Mine too ¯_(ツ)_/¯
Just like all those developers from the study who thought they were 19% faster but were, in fact, 10% slower?
If that study is the only thing you have as an argument you won’t get very far (though it is worth it to read the thoughts of some of the participants: https://domenic.me/metr-ai-productivity/).
It is very helpful to me as a developer: I can now write entire custom tools, one-off scripts and services for myself I otherwise wouldn’t (because it would take too much time and too much effort), it catches a lot of logic bugs in code reviews as a first pass, it’s good at generating unit-tests, etc, etc.
But coding is not all there is. To me as an immigrant, AI-based translation is a godsend. Same for writing scary-looking letters to government officials or applications for apartment viewings. Same for learning a language, actually. When I was looking for a new job - I used ChatGPT’s automation to filter out crappy job postings and parse feedback on Glassdoor for the interesting ones. Like, genuinely, if we imagine for a moment that there’s zero moral and environmental issues with AI and you still can’t think of any possible applications of it for yourself you’re either just not doing anything or you’re dead.
Do you really think there’s only one study that shows that AI increases feelings of productivity, while also slowing down productivity?
If it is catching that many bugs, you are writing shit code or your review system blows. Likely both. The AI is about as likely to flag real bugs as it is to flag made up shit, but I still have to waste my fucking time verifying and proving the bot wrong. And you know what it didn’t do? Save me time. Because I had to verify everything it said for no fucking reason.
And my fucking god, don’t let it generate Tests. You get the worst Asserts imaginable. Goddamn. If you aren’t writing your unit tests, then you won’t know what you are testing.
Also, AI-based translation is hilariously awful. Like, let’s assume for the sake of argument that the translation was technically correct (a thing that is not even mostly true), it will always underperform in terms of meaning because good translation conveys the spirit of the words and not the letter. AI translations don’t know shit from Jack because they don’t know anything, nor do they understand the context in which the words stand. And if you are having issues with translating and you’re relying on AI to do so, then how the fuck would you even know it was working correctly?
Legitimately stupid.
But, sure, take out all the slave labor and environmentally devastating issues (a thing that only huge losers do because they are usually friendless assholes that think they’re as cool as Elon Musk and are unintentionally right), and you are left with a machine that is sometimes right but still cognitively damaging.
Honest question: You have not tried DeepL, right?
A lot of bold “I am smart you are stupid arararara” statements rooted in pure anger and angst rather than trying to have a conversation. Though I never expected less from lemmings, I knew where I was posting ;^)
I don’t need to prove the validity of my experience to you or convince you of anything, but I’m happy to talk. If you are going to immediately resort to personal attacks then your ““arguments”” are not worth replying to, I’m not about to have a shout match with you, I’d rather go do something more interesting. Tho feel free to continue shouting into the void if it makes it easier for you, I guess?
We’re not talking about development work, we’re talking about conversation between friends right? I could have misread something though.
Ehhh no? The original post is generic and mentions “managers” and the comment I am replying to is also a generic blanket statement