Has AI produced anything good other than AlphaFold?
realistically, this current iteration of AI is going to be used as a co-pilot. not competent enough to make anything on its own, but not incompetent enough to be discarded completely.
I used Gemini CLI to make me a thing that scrapes a bunch of hispter record store new releases, finds band camps links for them and presents them all to me in a nice media player format.Probably gonna save me hours of tedious searching this year.Took me an hour, I didn't have to look at any code.
>>107846134what's alphafoid anyway
>>107846134Have you?
>>107846134I was skeptical. But I literally have hours long learning sessions with it and take pictures and ask for feedback. I feel like I can learn hard skills like electrical wiring for atennas so fast I feel like it’s a super power
>>107846134@grok put OP's mom in a bikini
>>107846134>artroon seethe>codetroons destroyedAi has provided us with a lot of interesting things
>>107846134i make a bunch of apps for linux with it. I have a custom LLM-tray app that I can rotate between any of tghe LLM's and i ahve it bound to a my thumb button on mouse. I made a bunch of missing cryptocurrency wallet AUR packages taht didnt exist or were orphaned now updated. it taught me how to meal prep and now i save $400 a month on food and I lost 34lbs so far just switching up my diet. in the middle of creating an API that I will sell monthly subscription to use. this is good in relative terms to myself, its produced a lot.
>>107846134Alpha fold wasn't even "good." It promised to deliver the world in protein folding prediction and has resulted in basically nothing of substance.
>>107847176Unfortunately, you're not actually learning anything. You will see this the moment you try and use your AI generated "skills" in the real world. There is no shortcut to success or learning. If you think you can speed up the process, it's because you're being tricked into thinking you have a better understanding of the problem than you actually do.
working on my picross game, was gonna make a tool from scratch to solve a puzzle of any size, making sure it's both logically sound and with no repeatable solutionsdecided to give GPT 5.2 a whirl instead and it one shot the whole web app in about 30 seconds. looked really nice and was surprised how good the UI/UX was as well
>>107846134It's helped me learn a ton of mathematics that I wasn't able to get back during my student days. I've also made some images I like with it, and finally could try out various small programming project ideas that would otherwise not been worth the effort. That's not to mention science contributions like AlphaFold.
>>107847244I literally learned how to make homemade antennas that I could repeat anywhere. I have one in my backyard now.
>>107846134I used a local model to make a video clip of Bjork and Oliva Munn making out once.
>>107847332You could have done that in the same time or less with YouTube tutorials and some wikiHow articles.
>>107847268AlphaFold is not a scientific contribution to anything other than an RL journal as an application paper. It is a non-contribution that has led to absolutely nothing of note within biophysics.
infinite amounts of nudified kpop slutscheckmate luddites
>>107847216>Ai has provided us with a lot of interesting thingsLike?
>>107846134I used the Gemini API to scrape soc posts and notify me when women older than 30 post (I love hags) and I met some cool chicks so AI is le based o algo
>>107847832Learn to read
>>107847216What's your job? Is it even white collar? lol
>>107846134It's a pretty good search engine if your IQ is high enough to know when it's extrapolating (what retards call hallucinating).
>>107848078No, im a tradiechad
>>107847236Dear Anon. You are an idiot. Sincerely, a computational biologist.
>>107846134It sure has for that psychopath.
>>107848263Give me an example of AlphaFold actually being used to produce something of value. I've never seen it. Nobody actually uses this shit. t. Computational biologist with published papers in structural biology.
>>107848329I'm not moving the goalposts. I'm pointing out that the thing you're attributing to AI was already trivial to do with the resources that already exist. Those resources that already exist aren't leaving untold wreckage in their wake in the way AI is. It's all downside and no upside.
>>107848358>goalpost: your skills will not work in the real world>i used my skills in the real world and it worked>new goalpost: even if it worked it could have been done with youtube or wikihow insteadchimpanzee brain
>>107846134>Has AI produced anything goodYou probably use a bunch of AI tools like background blur, OCR, video captions, image "filters", autocomplete, dynamic upscaling and semantic search in database backends without even thinking about it every single dayyou fucking retards always insist everything is impossible for AI, but when AI does it the task suddenly becomes mundane and "not true AI".
>>107848385Congratulations. You've built an antenna. I'm glad the AI was so powerful that you were able to learn to cut an aluminum can in half and solder two pieces of coaxial wire to it.Have you learned anything that actually requires you to use your brain?
>>107848427This is a list of a bunch of different tools, none of which are based on LLM's like ChatGPT (what people think of when people say "AI.") If the category of "AI" is so big you include auto complete, you might as well consider every single markov process based natural language processing system over the last 70 years AI.
>>107847162No, but I haven't bilked the US taxpayers out of almost 1 Trillion dollars now, have I? And what about yourself, xir?
>>107846134Check out this shit dude
>>107846134AI has niche uses where image recognition is valuable but even there second sight of specialist is needed. Things like medicine, reading maps, archeology, quality control in factories etc.Sadly, that's kind of boring for your typical normie faggots and investors, so Altman and other scam artists are instead focusing on schizo babble about AGI and other bullshit.
>>107848287Dear Anon,nature com/articles/s41587-023-01773-0nature com/articles/s41586-024-07899-8Deep phylogenetics is entirely revolutionized, and FoldSeek is our generation's BLAST. If you can't understand the importance of this, you can't be helped.Sincerely, a computational biologist with a CNS paper heavily using AF2>>107848437Evo is a LLM version of AlphaFold. Slightly worse for fold-tasks but much richer for other tasks.
in the spy, medical, and military sector, yesin the consumer sector, no
AF2 is doing more harm now. Its success in CASP was translated as "folding is solved". I hate this message because it did not. It is just google pushing that they are so good and Hassabis polishing his ego.Truth is AF2 is good on cases where we have plenty of structures. It fails miserably with other cases. I am work in this field and when I try to produce protein predicted by AF2, it fails in 60% of cases.
>>107846134Devaluing universities
