Rudolf Laine
709 posts
Rudolf Laine
@LRudL_
What I'm doing: https://t.co/7tVMLt1OwN For my writing, see: https://t.co/GwKY6jk3Tl
Katılım Ağustos 2022
414 Takip Edilen3K Takipçiler
everyone wants to be a moral patient but you need to be a moral doctor too
English
Rudolf Laine retweetledi
Sharing our work on full-duplex multimodal models -- real-time interaction that's natural and intuitive without compromising on intelligence.
We started Thinky in part to differentially advance capabilities for human-AI collaboration, which are underemphasized relative to intelligence/autonomy because they're harder to eval.
In the future, we think every AI system will have something like an interaction model as the outer user-facing layer, continually keeping the user informed and learning what they actually want.
Thinking Machines@thinkymachines
People talk, listen, watch, think, and collaborate at the same time, in real time. We've designed an AI that works with people the same way. We share our approach, early results, and a quick look at our model in action. thinkingmachines.ai/blog/interacti…
English
Rudolf Laine retweetledi
English
@ryancareyai The Intelligence Curse people wrote a scenario, it's called History of the Future, check it out!
English
Do worriers about "gradual disempowerment" have some concrete predictions so that we can do a similar accounting a few years from now?
Stefan Schubert@StefanFSchubert
AI propaganda has so far not been as effective as people feared a few years ago update.news/p/ai-economics…
English
In this world, after decades of gathering data, they'd fit a model to the drift that cuts the error down to centimeters, brag about how they improved accuracy six orders of magnitude and declare they "solved it with AI".
Andrew Gordon Wilson@andrewgwils
Without Einstein's general relativity from 1915, your GPS would drift about 10 km per day, and you'd have no idea why.
English
Working with Tinfoil was great. Trusted execution environments remain a very underrated technology for upholding privacy & freedom (for a hopefully-useful example of making them work at scale, see workshoplabs.ai/blog/private-p… )
Tanya Verma@t0nyav
Congrats @WorkshopLabs on joining @thinkymachines! We worked together on private post-training and their vision has strongly influenced our own at @TinfoilAI. At a time when so many feel disenfranchised by AI, @LRudL_ and @luke_drago_'s Intelligence Curse essays are some of the clearest writing on what's at stake, and their vision of AI that superpowers human agency is needed more than ever. Thinky is very lucky to have them.
English
Rudolf Laine retweetledi
As I've talked to folks about Thinky's purpose and why Thinky exists, I've found myself referring to The Intelligence Curse (intelligence-curse.ai) time and time again. So much so that I've even suggested in the past that The Intelligence Curse would serve pretty well as Thinky's mission statement :)
It's almost uncanny how much @luke_drago_ and @LRudL_ 's mission on keeping humans empowered matches @thinkymachines' mission. Excited to work together!
Workshop Labs@WorkshopLabs
Workshop Labs is joining @thinkymachines. We believe there's a path for AI to make humans matter more. We couldn’t be prouder to join Thinking Machines to see this work through. workshoplabs.ai/blog/wsl-joini…
English
Rudolf’s essays are what introduced me to AI safety in 2024, and I am lucky enough to know Luke personally. They are both extremely smart, articulate, capable individuals — this is tremendous news for the industry.
John Schulman@johnschulman2
Luke and Rudolf's writing on keeping humans central in an AI-powered world sparked a lot of discussion at Thinking Machines. For me, it captured some things I'd been thinking about but hadn't put as clearly. The more I got to know them and learned about their work, the more I wanted to work together. Really glad they're joining us.
English
I can’t talk enough about how extraordinary Rudolf is. N=1 researcher, and also somehow a philosopher, designer, and engineer. So glad we get to keep working together to keep the future human.
Rudolf Laine@LRudL_
At Workshop Labs, we dreamed we’d one day be a new type of frontier lab, focused on building AI that strengthens rather than degrades human relevance Turns out this already existed two months before @luke_drago_ & I published the Intelligence Curse essays: Thinking Machines
English
@brandonwilson @johnschulman2 I liked this essay! Recommend others read it. My thoughts at x.com/LRudL_/status/…
Rudolf Laine@LRudL_
Great essay that really understands how leverage, incentives, & human psychology drive history! Exceptionally clear-eyed and concise
English
@johnschulman2 @LRudL_ You should read my essay. I directly address their core thesis.
animalspirits.ink/the-leverage-t…
English
Rudolf Laine retweetledi
Luke and Rudolf's writing on keeping humans central in an AI-powered world sparked a lot of discussion at Thinking Machines. For me, it captured some things I'd been thinking about but hadn't put as clearly. The more I got to know them and learned about their work, the more I wanted to work together. Really glad they're joining us.
Workshop Labs@WorkshopLabs
Workshop Labs is joining @thinkymachines. We believe there's a path for AI to make humans matter more. We couldn’t be prouder to join Thinking Machines to see this work through. workshoplabs.ai/blog/wsl-joini…
English
Rudolf Laine retweetledi
Welcome @luke_drago_ and @LRudL_ to @thinkymachines. They started Workshop Labs determined to build AI that keeps the future human. They’ll continue that mission at Thinking Machines, where we create powerful AI systems that think alongside humans and extend our agency.
From Tinker to our research grants to the work we're doing to advance the frontier, everything we do is in service of the same mission -- AI that keeps our civilization empowered. Luke and Rudolf have been building toward the same thing.
There's a path for AI to make humans matter more. Glad to have them working on it with us.
Luke Drago@luke_drago_
Cat’s out of the bag! Today @WorkshopLabs is joining @ThinkyMachines. We started Workshop Labs to build towards a world where people still matter, even as powerful AI advances. I want to talk a bit about why this is the best move to serve that mission.
English
Even more, in our chats with @miramurati & @johnschulman2 & others at Thinky, we were also continuously heartened by their deep alignment on mission.
I think the mission goes much further if we work together. So we’re joining Thinky!
English
With our excellent small team, we built a lot: private post-training at SOTA open-source speeds, and a closed beta for a product around it.
But we also saw Thinky putting out research & infrastructure (@tinkerapi !) that we wished we could do
English
At Workshop Labs, we dreamed we’d one day be a new type of frontier lab, focused on building AI that strengthens rather than degrades human relevance
Turns out this already existed two months before @luke_drago_ & I published the Intelligence Curse essays: Thinking Machines
Workshop Labs@WorkshopLabs
Workshop Labs is joining @thinkymachines. We believe there's a path for AI to make humans matter more. We couldn’t be prouder to join Thinking Machines to see this work through. workshoplabs.ai/blog/wsl-joini…
English
@mentalgeorge @karpathy @WorkshopLabs Yes, needs training, but I suspect there are ways to go about it!
English
@LRudL_ @karpathy @WorkshopLabs Yeah the models are ridiculously uncalibrated on salience. My guess is this is something you need to train them to do explicitly, which is probably damn hard
English
One common issue with personalization in all LLMs is how distracting memory seems to be for the models. A single question from 2 months ago about some topic can keep coming up as some kind of a deep interest of mine with undue mentions in perpetuity. Some kind of trying too hard.
English
Yes you also need variation & differential success, but spread of trait being faster makes things go faster. E.g. a company’s practices are hard to copy because much is tacit and humans can only transfer that slowly; can easily add years to time of successful feature in some an industry spreading through that industry, even without moat-guarding. Or team member who discovers a new practice might take weeks to teach others. Of course, if this somehow completely collapses variance, that might slow things down on net!
English
@LRudL_ @TomDavidsonX Direct copying is inherently not going to increase cultural learning, which is obvious from Hayek
Agents need diverse context (and probably weights) to produce cultural learning
English
I think of cultural learning as a crucial part of the intelligence explosion, but don't think it could explode by itself
Historically, human cultural learning was super-exponential bc it was coupled with population growth and improvements in education. Without that, it would have fizzled out due to ideas getting harder to find
Similarly, for ai-driven cultural learning to be explosive it will need to be coupled with growth in the number of AI systems or their intelligence
Tyler John@tyler_m_john
I wrote an essay on the possibility of an intelligence explosion via cultural learning. Why are humans smart? Because we built a body of knowledge over 10,000 generations. This is the only form of intelligence explosion that's ever happened, so it could happen with AI too.
English