https://www.youtube.com/watch?v=1sd26pWhfmg>Large language models are now capable of automating attacks that were previously only possible by human adversaries. In this talk, I discuss several ways that adversaries could mis-use current models in order to cause harm both at a larger scale and at a lower cost than they do currently. For example, we find that recent state-of-the-art models can now find 0-day vulnerabilities in large software projects that have been extensively tested by humans for decades. These new capabilities will alter the threat landscape and require we rethink security in the coming years.https://x.com/deedydas/status/2038269993394356589>Legendary Don Knuth has now used AI to fully solve his Hamiltonian decomposition problem for odd and even cases.>Knuth: “We are living in very interesting times indeed.”It's so over. You can forget about employment. AI has now replaced even cyber security and research scientists.Human monkey is not needed anymore.
>>108480954Eh fuck it, maybe we can be good prompt engineers
>>108480954surely the billionaire class will take care of me and let me NEETmaxx when i become a disposable meatsack.... r-right...
>>108481036it's the Epstein class anonget with the times
>>108481036>"free healthcare is communism">"of course little code monkey, we will give you free money after you finish replacing yourself, trust us">>108481010they could just make a meta-LLM that's an expert prompter>>108480954honest question, how are companies okay with trusting their entire product on another company that uses all the data you feed it and your competitors also get access to it?When I worked in beg tech they were super annoying about supply chain attacks and shit like this
>>108480954>You can forget about employment.don't have to tell me twicein an ideal world I'd have my own homestead and have no reason to play the retarded rat race because property tax wouldn't exist
>>108480954Luddites lost. Aigods won
“Frontier models readily generate detailed image descriptions and elaborate reasoning traces, including pathology-biased clinical findings, for images never provided, we term this phenomenon mirage reasoning. Second, without any image input, models also attain strikingly high scores across general and medical multimodal benchmarks, bringing into question their utility and design. In the most extreme case, our model achieved the top rank on a standard chest X-ray question-answering benchmark without access to any images. "Benchmaxxingbros... i don't feel so good,
AI Chads won
>>108481284>how are companies okay with trusting their entire product on another company that uses all the data you feed it and your competitors also get access to it?companies already trust github and aws. why wouldn't they trust ai companies?
>>108481388>our model achieved the top rank on a standard chest X-ray question-answering benchmark without access to any imagesis this as retarded as I think it is?
>>108480954The four color problem was solved by computer in 1976.Only retards with limited long-term memory and short attention spans think there is anything of substance to this AI bullshit.
if you haven't found a twenty year old vulnerability in the linux kernel, you're not allowed to criticise or downplay claude's acheivements
>>108481728In fact, it was actually solved by a python program that was partially written by AI. Very similar to the computer’s exhaustive search on the four colour problem.
>>108480954Until it can generate me a girlfriend idgafChatbots don't count
>>108481010They've been researching orchestrator types, aka what the so called "prompt engineer" does, for a while already.
>>108480954If LLMs can do everything why haven't you used one to make you a millionaire yet? Are you retarded?
Funny coincidence
>>108481690yes the LLM hallucinates the input and the benchmark is so full of holes it works
Completely uselessAny grad student could sit down and do this in a couple of daysMeanwhile AI still can't do any useful
>>108480954it's funny watching people like knuth out themselves as pseuds
>>108480954why does ai struggle deriving the ground state for a 1d spin state then?
>>108483645retard loser
Idiots larping fantasy bullshitWake me up when llms can actually think
>>108480954GPT 5.4 thinking, having access to a bunch of mtg games and all of its rules, thinks a magic cards abilities are active when the card has left play. They're going to replace like 10% of the workforce and in a year this will be a disaster.
>>108481036They'll keep some of your kids around if they're pretty enough to rape
>>108480954>boomer on death's door thinks AI slop looks convincing This is really the crux of this entire bullshit industry on the sales side
>>108481418glad to see this meme is already getting run into the ground by people who don't understand even a single part of it
>>108483984You'll know this will have happened when OpenAI stops trying to sell tokens going in and out of ChatGPT and just starts building every successful global company but better
>>108485612you lost
It really is mind-blowing how the copy-paste machines with horrible efficiency in their copy-pasting trick retards over abd over again when they copy-paste something for the first time.
>>108481036Clearly. We are going to have to build our own society. They have theirs.
>>108485659Cope. Ai will replace you
>>108480954ITT: web shitters coping. Security and theory people are watching it in real time. You can literally get Opus 4.5 to scan advanced cryptography and security-related code and it will find actual math and logical bugs. The talk is 100% spot on, we're fucked. Right now you had something and asked an LLM to do the boring work to confirm or write the code, now the LLM will do the actual intellectual work and we'll have to confirm like a soulless industrial worker>>108482274Watch the talk, (smart) people are doing exactly that to find bugs in smart contracts and get millions. Also, cURL has been flooded with security reports so much that they closed the bounty program
>>108480954Serious question: Where do you pivot in CS to actually have some SOVL and not be a cog in the LLM machine feedback loop?
>>108486299nutrient paste to fuel the ai machine
>>108485659i wonder for how many years after the singularity you will be reading stale chud takes from 2022
>>108485659On the bright side, this scam has exposed how many people in positions of authority are retards.>>108486362The singularity of my cock and your ass?
>>108486288Answer the question, why aren't you exploiting OpenSSL right now and becoming a billionaire ?
