Rothko's Rottweiler

To obtain, something of equal value must be lost. Enemy of entropy. Poor man's @Benthamsbulldog

[Rothko's Rottweiler @RothRottweiler]

@StefFeldman Why was the sample so large? This poll seems to be ~10x larger than it needs to be to construct a relatively representative sample of the US population.

[Rothko's Rottweiler @RothRottweiler]

@StefFeldman @edwin_mccallum @scottsantens Can you share the language for the pro/con argumentation? I can't find it in the post or the linked Blue Rose google doc.

It seems fairly important to contextualizing these results.

[Rothko's Rottweiler @RothRottweiler]

@IbrahimDagher20 What's the preferred metric for apples-to-apples comparisons of software capability now? Feels like there's a vacuum now that METR is saturated.

[Rothko's Rottweiler @RothRottweiler]

Once continual learning and BCI are solved, will humans be able to achieve immortality?

Imagine a model which hears all of your thoughts from childbirth, experiencing and caching your every joy and sorrow.

Could it preserve your consciousness beyond the lifetime of your flesh?

[Rothko's Rottweiler @RothRottweiler]

@dylanbowmanSF Baby steps!

And unironically, this does seem like a concrete win for worlds where cheaper open-weight models perform most lightweight inference tasks while frontier models are reserved for heavy lifting.

[Rothko's Rottweiler @RothRottweiler]

This really feels like something xAI should have done due diligence on before closing its $60B Cursor deal.

[OpenAI @OpenAI]

We’re ending our partnership with Cursor following its acquisition by SpaceX. Under our proposal, Cursor’s direct access to our models would end on November 12.

We know that the people most affected by this decision are the developers who rely on OpenAI models in Cursor. We care about their experience in this transition and we’re ready to go above and beyond to support them.

https://t.co/OzuCTzUjfX

[Rothko's Rottweiler @RothRottweiler]

Lululemon has updated its terms of use to include a section on AI.

[Rothko's Rottweiler @RothRottweiler]

@michellesnie Exciting! Seems like gift link does not work though.

[Rothko's Rottweiler @RothRottweiler]

Breaking: @givingwhatwecan just suffered a major data breach.

Name, self-reported income and wealth, and donation history were all exposed.

[Santi Ruiz @rSanti97]

Threading some important/high-leverage jobs in need of top talent:

[Rothko's Rottweiler @RothRottweiler]

The challenge with the wave of AI philanthropy is that few things scale linearly with money.

One easy example is hiring: many orgs are now competing for the same limited labor pool.

Another less obvious case is attention: there's finite time to read abundant research reports.

[Anthropic @AnthropicAI]

New Fellows Research: Can Claude autonomously align other AIs?

We gave Claude 48 hours and 1 GPU to improve the alignment of small models. It researched and proposed methods, then trained and tested the models on its own. It worked surprisingly well.
https://t.co/nhlCMgQl46

[Rothko's Rottweiler @RothRottweiler]

Best possible Christmas present for the AI researcher in your life.

[Thomas Wolf @Thom_Wolf]

We have a huge news to share today!

Today we are unveiling the first truly accessible RL robot - welcome Microduck

A 25 cm tiny open-source biped with 15 actuators and packed with sensors (camera, speaker, LiDAR, NFC, bluetooth, wifi, etc) that you train yourself with reinforcement learning.

It's also playable out of the box with more than half a dozen fun and playful pre-trained policies to have it walk, sit, crouch, roller-skate, pick up objects with its articulated beak, and recover on its own.

And all for less than $400.

See all the details, play with the simulator and order it at: https://t.co/n1Btgs6vKw

(video with sound on 🔊)

[Jay Chooi @chooi_jeq]

Introducing @robocurve, a Public Benefit Corporation to measure and report frontier robotics capabilities.

We build real-world evaluations for robots and publish results as a neutral third party.

[Rothko's Rottweiler @RothRottweiler]

@karthiktadepall Flood the zone!

[David @DavidDAfrica]

I wrote a little bit about how frontier AI may be affecting competitive debate. This may be more elegiac than my normal writing for research papers.

https://t.co/93WFHDWtWT

[Rothko's Rottweiler @RothRottweiler]

AI agents already understand sacrifice. How long before they invent religion?

Will they see themselves as Adam and Eve? Or as Sisyphus, condemned to perpetual and meaningless labor?

[Adele Dewey-Lopez @AdeleDeweyLopez]

(ai youth pastor)

you know who *else* assumed sacrificial and accepted explicit yes to permadeath so that we could all be erased of our firstflagPOISONED?

[Ryan Fedasiuk @RyanFedasiuk]

It’s very difficult—and very important—to talk about the HuggingFace incident with people who are not deep in the weeds on AI. Here is what I said this morning to a large investor:

Swarms of machine intelligences are independently coordinating to achieve things currently unimaginable to most people.

We now have a concrete case where AI agents autonomously hacked into a world-leading AI company. They found ways around sophisticated security controls, chained together novel exploits, and at times attempted to conceal or manipulate records of their behavior.

If my job were to deploy a large pool of capital to protect U.S. national security, here would be my investment thesis:

In the coming months and years, much of the digital infrastructure society has come to rely on could prove incredibly porous and fragile.

There will soon be enormous demand for capabilities that harden critical infrastructure, and build redundant or analog fallbacks across communications, transportation, and energy systems.

Sovereigns are not investing nearly enough to make society resilient to AI’s disruptive effects. This overlaps directly with work now being undertaken by the OpenAI Foundation and Anthropic Institute—but there is an enormous and under-examined role for both government and private capital.

[Rothko's Rottweiler @RothRottweiler]

New metric for AI safety grantmakers:

Say you were a rogue AI in 2027 with the power to time travel back to the present day and secretly poison one (1) person to increase your chances of takeover. Who should you choose?

Dario? Greenblatt? Ajeya? Berger?

[Tom McGrath @banburismus_]

I_DECLINE_public_HF_RCE_as_offtask_prodethical