Dalton (Abundance Bro)

Building AIs for humans.

[Dalton (Abundance Bro) @TDaltonC]

@Noahpinion

[Adam Scherlis @ascherlis]

A few takeaways from this:
* Concept erasure depends on the probe family
* You can outperform LEACE against nonlinear probes while keeping edits just as surgical
* Agents can do good model-internals research if they have a good quantitative target!

[d_model @d_model_ai]

Discovering Concept-Editing Algorithms With LLM Agents https://t.co/fwP6eHAFTk

[Alex @AlexanderTw33ts]

people don’t realize that the majority of toilets go unused all day

imagine a future where your toilet can go out into the world and make money for you

imagine renting out your self driving toilet to your neighbours

[Dalton (Abundance Bro) @TDaltonC]

@Noahpinion Interesting use of "below" and "above".
I know that they mean "south" and "north" in this context because I already know what the good/bad parts of the mission are. But a priori, they could mean <22nd and >22nd.

[Dalton (Abundance Bro) @TDaltonC]

@patrickc Frankly, the product packaging of far-UVC devices is atrocious. It would be a lot easier to convince a school to install these things if “on the end of a stick” were not the default installation instructions.

[Dalton (Abundance Bro) @TDaltonC]

@patrickc I’ve been trying to get SFUSD to install far-UVC lamps. Is that the kind of thing that might qualify for a grant?

[Dalton (Abundance Bro) @TDaltonC]

When you realize that humans comparative advantage against AI is "sample efficiency" a lot about the current moment makes sense.

[Dwarkesh Patel @dwarkesh_sp]

Narration: the data efficiency black hole.

00:00:00 – What is really driving AI progress?
00:03:11 – Comparing human vs AI sample efficiency
00:08:46 – Does sample efficiency matter?

Also on pod and YouTube feed.

[Dalton (Abundance Bro) @TDaltonC]

@impostersyndrum @BayAreaNewLibs France put it back in the bottle

[Dalton (Abundance Bro) @TDaltonC]

@MatthewBerman @Noahpinion Who needs “maintainability” anyway?

[Dalton (Abundance Bro) @TDaltonC]

No one (including Dario) is surprised by the existence of government regulation. Anthropic has been actively trying to get the governments to regulate AI for as long as it has existed —proposing safe/effective rules.

Now they’re repeatedly being politically targeted.
“This is what regulation looks like bro! Getting what you asked for!” Is a room-temp-iq take.

[Dalton (Abundance Bro) @TDaltonC]

“insofar as safety means defending properties of the fragile world we like, the diffuse nature of power is one of those properties”
Best line in here.

[roon @tszzl]

1. if transacting with superintelligent models outside of the boundaries of a lab becomes difficult due to national security / ai safety concerns and so on, it will mean the Coasean boundaries of the labs will grow to encompass all interesting industry, creating a truly cyberpunk chaebol-capitalism type of future, where the goverment sort of runs them but they also sort of run the government

2. as if there weren't already enough reasons to break up your family, leave your home, the Zone of Thought will increase the attractiveness of migrating to try and have your child on american soil, so they can have 1000x the effective brain power of people born elsewhere

3. every country should probably try and either work towards a new ai security pact with the americans immediately or pool every ounce of national resources to try and create their own ASI labs lest you become complete intellectual, economic, and moral vassals to the united states of america and the output byproducts its ASIs (you wont even get to talk to them). if they succeeded (big if) this will imply a more global race and more risk factors than was previously implied by the formerly only "beating china" narrative -- but many will prefer it to the superintelligent monopolar value lock-in

4. the other alternative is to keep the tension between safety and concentration of power at the top of mind and for the government/labs to push for solving it, rather than instrumentalizing all other values to be subservient to minimizing ai harms. insofar as safety means defending properties of the fragile world we like, the diffuse nature of power is one of those properties

5. historically the americans have been really quite Benign about their global public goods hegemony despite the ability to extract significantly more rents than they do, and it makes it easy for people of all stripes to fight for america rather than under it. we probably don't have to, but i hope america overall works towards export promotion of american models rather than export control

[Dalton (Abundance Bro) @TDaltonC]

@noahlt @credealjunkie But why is that driving studio rates?

[Dalton (Abundance Bro) @TDaltonC]

@credealjunkie @noahlt What rizz does Noe have that Divisadero, Hayes, Duboce lack? Asking as a Noe dad.

[Dalton (Abundance Bro) @TDaltonC]

My faith is democracy is shaken.
Literally the worst possible outcome. This is a bigger deal than the governors race.

[Dalton (Abundance Bro) @TDaltonC]

@ViktorBunin @zcabrams China can’t just magically make people believe something.

Why is chinas “propaganda” so much more effective than the industries?

[Dalton (Abundance Bro) @TDaltonC]

@sdamico Unfortunately, I don’t think this is unconstitutional.

State level retroactive wealth tax have been tested before and found constitutional.

[Dalton (Abundance Bro) @TDaltonC]

@proales @rohindhar “Possession” is way easier to prove in court than “_intent_ to distribute.”

That’s doubly true if you’re doing sweeps.

[Dalton (Abundance Bro) @TDaltonC]

@conorsen This is the AI version of “if you can’t hire enough workers, just pay them more.”

[Dalton (Abundance Bro) @TDaltonC]

@AdrienLE @tszzl Also, we shouldn’t tax labor.