Gus Docker

Podcast Host @FLI_org

[Samuel Hammond 🦉 @hamandcheese]

Congress and the Judiciary need their own Blackwell clusters if only for constitutional, separations of power reasons.

[roon @tszzl]

just so everyone is clear: this is evil. you are justified in thinking it’s morally bad. tons of apologetics happening for bad people. if you think behavior like this is just desserts for the tech industry due to some hobbyhorse you have, you have gone insane

[Gus Docker @GusDocker]

@AaronBergman18 @jdcmedlock @AaronBergman18 can you expand on 3? Link to any writings on the subject?

[Benjamin Todd @ben_j_todd]

Why do we have to keep doing this. We have AGI (in the sense that matters) when you can hire an AI to do 95%+ of jobs, or fully automate AI R&D or scientific research.

Can you hire an AI to replace yourself? No? Then we don't have AGI.

[Future of Life Institute @FLI_org]

"There is no gain from getting to superintelligence. The only actor gaining is the superintelligence itself."

-@ControlAI founder and CEO @AndreaMiotti on the latest FLI Podcast episode with host @GusDocker, available now at the link or on your favorite podcast player!

⬇️ 🔗

[Future of Life Institute @FLI_org]

@ControlAI @andreamiotti @GusDocker https://t.co/zoVuYIFaib

[Anthony Aguirre @AnthonyNAguirre]

When people think AI shouldn't replace people in doing something, it would be much more useful for them to just say so, instead of declaring that it can't.
There is nothing wrong with reserving certain things – indeed a great many things – for humans to do. But pretending that AI can't be stuck into a human's place, whether effectively or not, does not help us do so.

[Future of Life Institute @FLI_org]

🚨 "We have seen significant periods of instability, unrest, even revolution with inequality and when the gains from the economy are not well distributed enough, and we might see similar outcomes if we see that the wealth [from AI] gets overly concentrated within a small set of people."

📻 New on the FLI Podcast: @WindfallTrust Director of Research Deric Cheng joins @GusDocker to discuss how AI could reshape the social contract and global economy.

🔗 Listen now at the link in the replies:

[Future of Life Institute @FLI_org]

🆕 📻 On the latest FLI Podcast episode, Future of Life Foundation researcher Oly Sourbut joins host @GusDocker to discuss how AI could help humans reason better.

🔗 Watch in full at the link below:

[Future of Life Institute @FLI_org]

"I'd rather get all the flourishing things out of not superintelligent systems but highly capable systems that I can coordinate with well & that can coordinate with each other well, as opposed to training successor agents that we don't know how to train."

-@AmmannNora on the FLI Podcast:

[Nora Ammann @AmmannNora]

Had the pleasure to speak to @GusDocker at the @FLI_org podcast about AI - what I'm worried about, but also what I think we can do to help things go better!

We touch on...🧵

[Future of Life Institute @FLI_org]

🚨 "Why are companies building these things? The REAL reason, the goal, is to not give people the tools that will just make them more productive, but to replace people."

-@AnthonyNAguirre on the FLI Podcast with host @GusDocker ⬇️ 🎥

[Future of Life Institute @FLI_org]

🆕"If the final input at the end of the day that informs regulation is what the public wants and who they vote for, then at some point the money stops working for you."

-@TheMidasProj's @tyler_johnston on the FLI Podcast w/ @GusDocker, discussing how to hold Big AI accountable 🔗👇

[Future of Life Institute @FLI_org]

"Better futures - namely, trying to make the future better, conditional on there being no catastrophe - is in at least the same ballpark of priority as reducing existential risk itself."

New on the FLI Podcast, Forethought senior research fellow @willmacaskill joins host @GusDocker to discuss the contents of his new essay series, "Better Futures".

🔗 Listen to the full episode at the link in the replies below:

[Future of Life Institute @FLI_org]

"If you feel like, 'hey, we're actually not hitting certain alignment things right now and we're using misaligned models to try and align models of the future'... probably good to speak up now."

-Karl Koch, founder of the AI Whistleblower Initiative @AIWI_Official, on the latest FLI Podcast episode with host @GusDocker.

🔗 Listen to the full episode at the link in the replies:

[Rob Wiblin @robertwiblin]

I think the complaints that such statements are not specific or enforceable enough are misguided — if widely supported internationally it could be made specific and enforced well enough.

But Dean makes a substantive argument that this broad path would naturally lead to a government monopoly on superintelligence, and that's actually riskier, all things considered, than a decentralised/chaotic AI rollout.

I don't agree, but it's not a stupid argument, and the balance of risk will come down to very challenging guesses about the difficulty of technical alignment, the ease of bioweapon development, how the government would use AGI, how AI would be adapted in the military, and on and on.

In a way Dean is really pointing out that our situation is even scarier and more precarious than you might otherwise think, because unfortunately there's no actor we can trust not to act self-servingly.

[Future of Life Institute @FLI_org]

📻 "If you're a business, you wanna make money, you wanna chase profits, you have shareholders, fine.
What actually irks me personally is when people try to have it both ways, in the way that the leaders of OpenAI do, where they try and speak as if they're still a nonprofit who are doing things for the benefit of humanity... and they're clearly not."

🗣️ @business tech columnist and "Supremacy" author @Parmy discussing how AI companies have transitioned from research labs to product-led businesses, on the latest FLI Podcast episode 🔗👇

[Future of Life Institute @FLI_org]

Listen to @Parmy's full conversation with host @GusDocker: https://t.co/bNNh5WOWO9

[Future of Life Institute @FLI_org]

🤖 "If you have an agent that has very broad goals and a very open-ended autonomy, you're gonna lose a lot of meaningful oversight of that system, most likely. So, that's the biggest shift I think between a Tool AI and this more agentic path that we're on right now. I think you could have a Tool AI that's still an agent, but it would have a very bounded autonomy."

-📻 @ForesightInst @HopeExistential Program Director Beatrice Erkers on the newest FLI episode with host @GusDocker.

🔗 Tune in now: https://t.co/ZtzWyaI6nw

[Future of Life Institute @FLI_org]

🆕 "As we continue to build technology that is designed to replace rather than to augment, we move closer and closer towards a world where people just don't matter. And then of course you're reliant on other forces [...] it's a very precarious situation to be in."

-@luke_drago_ (co-author of "The Intelligence Curse" essay series; @WorkshopLabsPBC co-founder) with host @GusDocker on the newest FLI Podcast episode.

📻 Listen now at the link in the replies: