Skip Navigation

Posts
2
Comments
438
Joined
3 yr. ago

  • They made hundreds and 16 worked (when synthesized by humans), seems stochastic to me. Humans do better to my knowledge (I'm out of touch but do have a relevant background). It's based on LLM architecture (using genes and proteins in place of language), also known as a stochastic parrot, which was the sense in which I was using it.

    I don't deny it's a scary technology, more because of the likelihood of mistakes (or as AI companies like to say, 'hallucinations', anthropomorphising as advertising) than any intention.

    As usual with this type of technology you need to be competent with the field before it becomes a force multiplier, and those people could already make viruses, and have been able to for a long time. For example, one of the previous stochastic methods was randomized mutation in a specific environment.

  • I'd say it'd take a fair bit more, but it's a risk, hence

    double down on the manufacturing security

  • We've been living with the human capability to make and deliver super lethal viruses for what 40+ years and have methods in place to control it at the manufacturing level. 'genAI' is way too incompetent at this stage to achieve both a functional virus and evade those manufacturing safeguards and likely LLMs always will be, too stochastic. This is quite aligned to look how scary our AI is marketing.

    That said, virus is one of the studied extinction event scenarios for a true AGI/ASI. It would be prudent to double down on the manufacturing security, or, you know, go Butlerian at least until we have an excellent grasp on alignment (IMO you can't align an ASI, it's literally beyond our comprehension). But capitalism's gonna capitalist, existential threat be damned, just look at global warming.

  • It's not. There's plenty of time, chill and enjoy the ride.

  • Tossup between teaching geneticists not to use dumb statistics and a couple of minor kernel contributions along the way. Back after the first human genome everything was showing significant correlations to brain proteins, which are notably longer than somatic ones, because they were failing to normalize against length. Made a web app that did it for them, now it's part of the standard analysis packages. Not sure how long that would have kept going without my publication, but I got us there faster, so science progressed faster.

  • Fringe is my ideal sci-fi show: weird, far reaching, and is allowed to reach its end.

    Top 5 for me too, but I'd argue it (and Person of Interest) had unfortunately rushed final seasons. Still, gloriously weird, top notch acting, class act writing with some standout moments of utter glory (white tulip, that dandelion poking out to "Only You", my very favourite thing…). Hmm, shall bump it in my rewatch queue.

  • Presumably this has been quietly going along while Musk's neurolink shitshow has been getting the headlines. Here's to responsible usage of BCIs like this, and never letting whatever neurolink is and wants to become (seems torment nexusy to me) anywhere near any of our brains.

  • They Live

    "I have come here to chew bubblegum and kick ass... and I'm all out of bubblegum."

  • Super cool use of lenticulars.

  • Doing God's work ! (says the agnostic). This will be such a big win for the fediverse.

  • brain anthropy

    case in point ;)

  • Douglas Adams. Came across a signing and by random had borrowed the hardcover of So Long and Thanks For All the Fish (with the hologram on the front) and took it as a sign. Paid off the library...

  • Or, people's retirement funds via IPO. If you can change your plans investment strategy for your good and societies.

  • it’s what Ollama uses under the hood

    Actually they have their own, shittier engine now. They didn't credit llama.cpp while using that for quite a while. Agree with don't use if at all possible (unfortunately some third party softwares need it, good sign to use something else, but needs must where the devil drives)

  • It's unstable / explosive. (Actually TNT is pretty damn stable for an explosive, nitroglycerin would have been a better choice)

  • True, but not hugely relevant to models fitting in 32GB as I understand it, more for improving latencies in larger models (120B+) or batch processing. Nonetheless layer will work for OPs use case.

  • Inference doesn't actually need gobs of PCIe bandwidth, the model gets split in half at the layer level and you just pass the output of one card to the input of the next. I have 2x7800XTs going and one is on (chipset) PCIe4x4 and it works fine. External via eGPU can also be done.

  • It already is, this beats the NSA and OpenAI et. al.

  • At it's best you're paying (or better, your country is paying) to learn how to learn, with a specific discipline as a workspace. As with so much else it has enshittified due to rampant greed and power's fear of a well educated populace, still, if you go into it with that goal, some vestige can still be had.

  • Technology @lemmy.world

    The great decoupling: How we became the Eloi

    independentaustralia.net /politics/politics-display/the-great-decoupling-how-we-became-the-eloi,21100
  • Technology @lemmy.world

    Sam Altman’s Dirty DRAM Deal