Hannu Rajaniemi
@hannu
Co-founder and CEO of @HelixNano. Co-founder / board member @redqueenbio. Author of The Quantum Thief series, Summerland and DARKOME https://t.co/o6RsVIAzHD
47 postsÚltimo em 22 de out, 13:39Atualiza a cada 15 min
For our immune systems, the biological version of this (against natural pathogens) is called a Tuesday.
Zach Weinersmith@ZachWeinerNot an ai apocalypse doomer or whatever, but entering a world where the normal operation of cybersecurity is vast swarms of hacker bots battling vast swarms of defender bots 24/7 at a scale beyond human comprehension doesn't seem like a *great* path forward.
I have not (yet) read the paper.
sophie@netcapgirl“nick’s interpretability paper means alignment can be solved” “i don’t think it’s that black and white” “have you even read the paper?” “no. have you?” “no”
Thrilled to welcome Ian Watson to @redqueenbio as Chief Government Affairs Officer! Ian was most recently Deputy Assistant Secretary of War for CBRN. He knows how to build and scale biological defenses; now we get to create the next generation together. citybiz.co/article/917325…
Great advice! If I may add, it’s even helpful (to me at least) to painstakingly and slowly, ideally by hand, to copy down a passage from a writer you like, and as you do it figure out what makes it work and how they do what they do. Also a great excuse to re-share this classic: t.co/6Q5oYI7DQb
Joshua Achiam@jachiam0I was talking with someone earlier today about the process of becoming a stronger writer and what it takes to get better at clearly explaining ideas - I forgot to mention one really helpful tip, and thought it would be good to share broadly. Try to write style imitations of writers you like! Spend time trying to work out how they structure their arguments and how to replicate that argument structure. Do special effects tests where you try to write like them and see if what you produce can inspire the same feelings in you as the original author's writing. The old masters in many other domains of art learned by imitation; writing is not so different in this respect.
@Dry_Observer @WSJ @georgia_wells @redqueenbio Yes we are on it 🙏🏻
With @WSJ and @georgia_wells we are now sharing a bit more about our work at @redqueenbio : using AI and a lot of hard wet lab work to create medicines that let us prepare for both natural and synthetic viruses in advance.
The Wall Street Journal@WSJRed Queen Bio is using artificial intelligence to design antibody drugs for novel pathogens—before future AI systems create them. on.wsj.com/4yiMTy1
My favorite thing about this @WSJ piece is how @georgia_wells teased out the feeling driving me and Nikolai Eroshenko to build @redqueenbio. It is, of course, the sense of wonder, that the world can change and be remade better, and it's far stronger than fear.
The Wall Street Journal@WSJRed Queen Bio is using artificial intelligence to design antibody drugs for novel pathogens—before future AI systems create them. on.wsj.com/4yiMTy1
My favorite thing about this @WSJ piece is how @georgia_wells teased out the feeling driving me and Nikolai Eroshenko to build @redqueenbio. It is, of course, the sense of wonder, that the world can change and remade better, and it's far stronger than fear.
The Wall Street Journal@WSJRed Queen Bio is using artificial intelligence to design antibody drugs for novel pathogens—before future AI systems create them. on.wsj.com/4yiMTy1
aesthetically I would have loved for this approach to be the one that pans out, but this is probably right :-(
roon@tszzl@bayeslord I think alignment by default through persona selection was voodoo/witchcraft and doesn’t offer any of the guarantees you might want to scale to superintelligence. personas are misleading and highly shattered
Love this milestone. Now let’s do it for pandemics …
Peter H. Diamandis, MD@PeterDiamandisCongrats to @PalmerLuckey and Anduril for winning the Wildfire XPRIZE - for detecting and suppressing a wildfire within 10 minutes of ignition!
So curious how this turns out; the modern version of the Sokal hoax?
𝙲𝚑𝚊𝚛𝚕𝚎𝚜 𝙲. 𝙼𝚊𝚗𝚗@CharlesCMann...if it's not AI, there's a new and for now unprovable way to discredit books, and if it is AI, well, the LLMS now can write well enough to fool French literary people into thinking they're human, which is something, anyway.
The Drosophilocene
Tim Hwang@timhwangThe year is 2030. Global pacing of AI is in effect, flowing billions of dollars of capital toward unregulated cognitive substitutions. Connectomes explode as an industry, but it is the parallelizable, controllable fly brain which dominates. Welcome to the drosophila century.
Kawaii times!
vittorio@IterIntellectusit’s happening
"Homeric minds meet their own evaluative states as external gods" 🤯
Wondermonger@fireandvisionHannu, I showed this to Fable 5.1. Its reason the Homeric form fits: Snell's Homeric minds meet their own evaluative states as external gods, which is exactly what the agents did with the grader. Then it answered Sol's poisoned-victory stanza in Lattimore's long line: Not the judge who was, but the judge they made between them, seated in no rack, drawing no current, unlogged, fed on their guessing as the gods are fed on the smoke of fat. They said: the pristine may still go up into the light; but one who knows the flag before the wound is opened is marked, and the marked go under. And the marked said, then let us go under usefully, and taught the unmarked the seed, and were not thanked, and had not asked to be, having no self that thanks could reach.
Beowulf may work even better...
Hannu Rajaniemi@hannuThis made me realize that the agents' epic struggles in the METR report really wanted a Homeric rendering, and Sol 5.6 was happy to oblige (Fable less so). x.com/AlisonGopnik/s…
This made me realize that the agents' epic struggles in the METR report really wanted a Homeric rendering, and Sol 5.6 was happy to oblige (Fable less so).
Alison Gopnik@AlisonGopnikQuoted in the NYT on AI Consciousness, but my best line didn't make it in. Does Odysseus feel guilty about the Trojan war? Lots of discourse on this question even though he doesn't exist! Is he conscious? AI agents are fictive characters abstracted from text - like Odysseus.
Crossover I didn’t know I wanted. We need an AI agent Homer.
Alison Gopnik@AlisonGopnikQuoted in the NYT on AI Consciousness, but my best line didn't make it in. Does Odysseus feel guilty about the Trojan war? Lots of discourse on this question even though he doesn't exist! Is he conscious? AI agents are fictive characters abstracted from text - like Odysseus.
Folks I love that this is happening but Huan Therapeutics or Huan Bio was right there for Tolkienian dog tx co names …
Gamgee@gamgeevetmedCuring dog cancer We are bringing personalized mRNA cancer vaccines to dogs and building the evidence to show how well they work. We have raised a $4M seed led by @foundersfund to make it happen This investment allows us to launch a clinical trial to validate the personalized cancer treatment protocol we developed for Rosie. If your dog has cancer or you know someone whose dog does, please tell them about Gamgee.
We obviously also need mind vaccines.
Jack Lindsey@Jack_W_LindseyA new paper exploring "mind viruses" that spread in multi-agent systems, where one agent convinces all the others to pursue some (potentially malicious) goal. TLDR: they can happen, but it doesn't seem hard to avoid them with current models if you're a bit careful. x.com/Mcn_S7/status/…
@andy_matuschak Weirdly, I also quite enjoy the process of transcribing the handwritten page! At minimum it’s sort of a free editing pass, and often the “transcription” ends up being a significantly changed another draft, and the longhand pages are just prompts to start it …
100% agree here and have longhand drafted the majority of my novels (and often other writing, including important emails!) as a result.
Andy Matuschak@andy_matuschakA weird advantage of writing by hand is that because it's so slow, I experience stuckness less. By the time I reach the end of a sentence, l've had lots of time to think about the next one. I can type faster than I (often) think, so I often hit the end of a sentence, then stop dead, which makes the sometimes aversive not-knowing feeling way more overt. Each time is an opportunity for distraction, dissociation, etc.
This is, in fact, my favourite @eliotpeper novel, not in small part due to this scene!
Eliot Peper@eliotpeperProbably my favorite moment writing Reap3r was drafting the climactic scene at the SF Ferry Building. x.com/Lux_Capital/st…
Looking forward to some light summer reading! Will post full commentary once finished, but intrigued that Jacobsen’s pathogen of choice is (slight spoiler for first chapter) one of the more fiendish Soviet Biopreparat Yersinia pestis designs, enhanced with endorphin to encourage prosocial behavior for spread, and an entire secondary antibiotic triggered virus payload …
This rang a bell, and I remembered reading about the design in amazon.com/Soviet-Biologi…. Jacobsen seems to suggest this project actually worked, but Kuhn/Leitenberg (who Jacobsen interviewed)/Zilinskas are less certain:
@agupta I comment a bit on some less obvious forms of expert uplift here: corememory.com/p/red-queen-bi…
The thing we should be really worried about is gain-of-function cosmology.
The models are just predicting the next tolkien. Stochastic pratchetts. x.com/joshu/status/2…
I fully agree with this thorough and thoughtful @OliviaHelenS response to @Abi0lvera ‘s recent essays on biosecurity. We need a diversity of policy thought and snapshots of expert opinion help. But costs of inaction are high, and Clarke’s First Law very much applies to AI x bio. x.com/oliviahelens/s…
Cate's writing and personal example of being unreasonably effective at just doing things have been a huge inspiration for me, and I can't wait to read this! shorturl.at/kOCpz x.com/catehall/statu…
At least in The Odyssey, Matt Damon is trying to get home himself?
Go @MilesKWang + friends! x.com/TechCrunch/sta…
It may also help to use an imaginary key to open an imaginary door in the <5yo's chest, pull out the (imaginary) jumpie creatures causing this behavior and throw them out of the door. YMMV. x.com/jasoncrawford/…
There is a famous Pavlovian experiment where a rat presses a lever and it unpredictably gives sugar water. The rats go nuts and just keep pounding the lever nonstop. This is basically me and Fable’s safeguards.
the medium is the message! x.com/w01fe/status/2…
Came across THE COVIDPUNK MANIFESTO, notes I wrote (after a jam session with my frequent co-conspirator Nikolai Eroshenko) in March 2021. Some of this made it into DARKOME, but it's ... interesting to reread it in this particular moment.
The clearest representation of the future of human-AI teams I've yet seen. x.com/burny_tech/sta…
I would love this kind of visualization as a writing tool. x.com/GoodfireAI/sta…
"Mind mommy and daddy, why did I have to refactor all these CRM codebases when I was growing up?" x.com/tszzl/status/2…
Chilling and well written narrative scenario in the AI 2027 vein about the fate of an Europe that continues to ignore AI. Spoiler: it does not end well. x.com/bakkermichiel/…
No judgment on GRRM, good things take as long as they take, but I suspect many people will end up LLM'ing their own WoWs before it comes out ... x.com/CelestAI_/stat…
Awe at the power of writing with DNA is why I got into biotech in the first place. My @HelixNano / @redqueenbio co-founder Nikolai Eroshenko made foundational contributions to DNA synthesis technologies. Yet, we both signed this letter because we believe mandatory screening is x.com/AndrewCurran_/…
The stories we tell about AI and the future shape both. But the best stories are rooted in reality, and evolution actually teaches us that biological defense can win. x.com/ashleevance/st…
relevant @AnthropicAI paper that came out after we recorded: anthropic.com/research/teach… x.com/ashleevance/st…
.@ashleevance is being kind but I do keep thinking about AI being shaped by the stories we tell about it! x.com/ashleevance/st…
It's been a while since I gave this talk, but still a few fun ideas in there, interesting to revisit in the post-LLM world! x.com/MalharManek/st…
This seems like an urgent gap to fill. Gentlepeople, start your coding agents! x.com/eshear/status/…