Rudolf Laine

What I'm doing: my writing, see:

[Rudolf Laine @LRudL_]

@GuiveAssadi

[Rudolf Laine @LRudL_]

THE COURSE OF TECHNOLOGY IS NOT INEVITABLE, ITS DIRECTION IS A CHOICE. Build the good thing!

[Anton Leicht @anton_d_leicht]

i still think we're in an automation-augmentation race: whether the future features fulfilling human work isn't something we discover, but depends on technical and policy work today.

if you're looking to join the race on the side of augmentation, this is a great place to start.

[Rudolf Laine @LRudL_]

Most ages of invention are also driven by (and end with the decline of) a specific social institution staffed by a specific Type of Guy that is structured to solve certain kinds of problem and depends on hyperstition. E.g. the Baconian program and the early modern scientific community of letters, the early-1800s amateur inventors, the late 1800s Edison/Tesla -style lab, the military-industrial state-sponsored WW2 research effort, Bell Labs, the ad-funded big-tech company, the SaaS startup, and the frontier lab. They are not interchangeable in what they pursue or who runs them or what mission they serve. It's not a Civilization game, there is no impartial "more progress" lever that is cranked repeatedly to roll along a pre-built track. The hyperstitioning means it's often hard to cancel one once it's launched though, of course

[Rudolf Laine @LRudL_]

@GuiveAssadi Consider also: LLMs product space is vastly unexplored (and exploration is mostly contingent on what people decide to explore) and choices like emphasizing chat early on or defining progress as autonomy (vs e.g. productivity uplift) will have long-lasting effects

[Rudolf Laine @LRudL_]

@luke_drago_ UNDERRATED TRUTHS NEED TO BE BROADCAST LOUDLY

[Rudolf Laine @LRudL_]

@captgouda24 i always write until i reach a fixed point of the lossless semantic summarization operation

but then people ask me why my ai scenario series is 20k words or whatever

i explain it says many things

“but surely you could’ve said it shorter”

no. read it again.

[Rudolf Laine @LRudL_]

@austinc3301 Ahem

[Rudolf Laine @LRudL_]

Q: how do I apply?
A: @MATSprogram has the link, it’s listed under “strategy and forecasting”

[Rudolf Laine @LRudL_]

@luke_drago_ & I will be running a stream at @MATSprogram this fall on issues like the intelligence curse, gradual disempowerment.

In the 1800s, people figured out how to orient society so that industrialization was good for people. We need this for AI

FAQs:

[Rudolf Laine @LRudL_]

Q: what sort of people should apply?
A: people with deep unease about existing ideas & paradigms here, who have experience in one or more of econ, history, policy/gov, AI futurism/forecasting, or political philosophy

[Rudolf Laine @LRudL_]

Sufficiently aligned AI feels like tool AI

[Rudolf Laine @LRudL_]

everyone wants to be a moral patient but you need to be a moral doctor too

[John Schulman @johnschulman2]

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.

[alth0u🧶 @alth0u]

the only choice a soul can make in the face of any technology that does not structurally appeal to our better natures but instead treats us as temporary scaffolding, conduits, and passthrus to be replaced

is to disengage and build another option for self and kind

[Rudolf Laine @LRudL_]

@DKokotajlo @ryancareyai Can confirm, it’s here: https://t.co/ZVEXsTD29j

[Rudolf Laine @LRudL_]

@pfau Yep! From https://t.co/Jz2UyVYp0V :

[Rudolf Laine @LRudL_]

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 https://t.co/uA5FTZxEgG )

[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.