Adrià 🌲🌳

\PhD student in Neurorobotics
\effective human optimism

[Adrià 🌲🌳 @AdrMmp]

Me enamoro. Motivación para estudiar ciencias y hacer posible esto algún día. Entre otras cosas. https://t.co/6MHUPltRYY

[Adrià 🌲🌳 @AdrMmp]

@kimmonismus but Ilya said in the Dwarkesh interview that they may release intermediate models, and remarked it a couple of times (43:09, 56:12)
https://t.co/d9z20y8DdF

[Adrià 🌲🌳 @AdrMmp]

@kimmonismus hopefully not LLM, would be cool to see something fresh

[Adrià 🌲🌳 @AdrMmp]

@thsottiaux Codex since 5.5 has been my main working model for everything, very thorough, rarely misses anything, even finds legit stuff to improve I didn't ask for. 5.6 Sol just does that with even better reliability and ease to deal with complex stuff like it's nothing

[Adrià 🌲🌳 @AdrMmp]

@DotCSV El problema con Fable/Mythos es que no es "la humanidad", es un solo lab y los partners que ellos consideren adecuados, dando al resto una versión nerfeada y con filtros de seguridad que lo hacen inusable para uso científico.
Veremos cómo evoluciona esto...

[Adrià 🌲🌳 @AdrMmp]

@DeryaTR_ I also asked a simple question about my research on spiking neural networks, and instantly got rejected, wtf is this

[Adrià 🌲🌳 @AdrMmp]

@OfficialLoganK @HCSolakoglu is the "Extended" in the Gemini app the same the same as "High" in aistudio?
@OfficialLoganK

[Adrià 🌲🌳 @AdrMmp]

@OfficialLoganK @HCSolakoglu make it like chatgpt, low thinking effort by default and the option to set higher effort for those who want, the majority will not change it as seen in chatgpt

[Adrià 🌲🌳 @AdrMmp]

@pdxweb @OfficialLoganK @HCSolakoglu true, but I am not sure the extended there is equivalent to aistudio's, it's a bit confusing.
haven't tested enough lately but gemini's had a history of being worse in the app.

if now's the same, then yes that would be it

[Adrià 🌲🌳 @AdrMmp]

@kimmonismus I've lived there for 7 years, missing it. You should definitely go

[Adrià 🌲🌳 @AdrMmp]

@IterIntellectus the relational sector, where the human element itself is part of the value (arts, education, care, entertainment, etc)
https://t.co/PMGHxt8u29

[Alex Imas @alexolegimas]

New essay on the economics of structural change and the post-commodity future of work.

1. Almost any question about the impact of advanced AI on the economy needs to start at the same place: what is still scarce? Answer that, and the analysis becomes pretty straightforward. This essay explores what becomes scarce if AI really can replicate most of what humans do in production, and what this mean for the future of jobs.

2. My conjecture, working through the economics: labor reallocates across sectors, and the sector it reallocates to has properties that keep labor a meaningful share of the economy. Ultimately this is about the structure of demand itself. For this, we have to go back to Girard, Augustine and Rousseau: once people's base needs are met, their preferences shift to comparative motives (e.g., status, exclusivity, social desirability). This motive is inherently non-satiated.

4. The key paper is Comin, Lashkari, and Mestieri (Econometrica 2021). As people get richer, they don't buy proportionally more of everything. They shift spending toward sectors with higher income elasticity. They estimate income effects account for 75%+ of observed structural change.

5. The ironic consequence: the sector that gets automated becomes a smaller share of the economy, not a larger one. Agriculture got massively more productive and its share of employment collapsed. Manufacturing too. The "stagnant" sectors absorb the spending and the jobs.

6. So the question is: which sectors have high income elasticity in a post-AGI world? I argue it's what I call the relational sector. Categories where the human isn't just an input into production, it is part of the value.

7. Why does the relational sector have high income elasticity? Because human desire has a mimetic, relational dimension. We don't just want things for their intrinsic properties. We want what others want, and we want it more when others can't have it. Girard, Rousseau, Augustine, and Hobbes all saw this.

8. In work with Kristóf Madarász, we showed this experimentally: WTP roughly doubles when a random subset of others is excluded from the good. And in new work with Graelin Mandel, AI involvement kills the premium. Human-made art gains 44% from exclusivity; AI-made art only 21%.

9. This all comes together for the core argument. The sector that absorbs spending as AI makes commodity production cheap is one where human provenance is part of the value, and demand for it grows faster than income. Exactly the profile that keeps labor meaningful.

10. To be clear about the claim: I'm NOT saying aggregate labor share must rise. It may fall. The claim is about sectoral composition, i.e., where expenditure and employment go once commodities get cheap, and the fact that the sector that will absorb reallocated labor maps to a substantial component of human preferences and desire.

11. If you're interested in the formal model, a linked companion technical note works out all the economics.

Read the essay here: https://t.co/NcjVgn2o8g

[Adrià 🌲🌳 @AdrMmp]

@VictorTaelin I've resorted to chrome plugins and html parsing scripts just for this...

[Adrià 🌲🌳 @AdrMmp]

humanity alignment's percentage, we need to increase it over the coming decades

[prerat @prerat]

wtf is 58% a law of nature or something

[Adrià 🌲🌳 @AdrMmp]

@prerat it's the percentage of humanity alignment

[Adrià 🌲🌳 @AdrMmp]

@DarthDromos @steelmansteveo @cremieuxrecueil Ok but you see, now it's easier to qualify and judge people.
In the blender you're stupid for jumping, "why the f would you enter an experiment where you can die?"; in the blue-default, you're evil, "why the f would you enter an experiment where you can kill people?"

[Adrià 🌲🌳 @AdrMmp]

@DarthDromos @steelmansteveo @cremieuxrecueil Yes exactly, but you can say the same for my version in the previous picture, where you're team blue by default and not pressing red skips the situation.

Just to illustrate that depending on how it's framed it can bias the choice.

[Adrià 🌲🌳 @AdrMmp]

@steelmansteveo @cremieuxrecueil Yes I think my version is more a counter to this version where red is inaction
https://t.co/t1QmUYsjr4

[Roko 🐉 @RokoMijic]

We're doing the "Blender" game again

There is a large blender.

Everyone in the world has to decide whether to step into the blender. If at least 50% of the people do step into the blender, it will be unable to overcome their inertia to get started, and everyone survives.

If less than 50% of the people step into the blender, then they all get blended up into paste and die.

People who do not step into the blender suffer no adverse effects.

Would you step into the blender?

(Blue=step into the blender, Red= don't do that)

[Adrià 🌲🌳 @AdrMmp]

So again, this was all a bet on that culture, to some extent I would sacrifice myself to keep that part of civilization going (if I assume that loosing like-minded people comprising nearly half the population could lead to a world where that civilization is lost) and making more likely its values prevail and keep improving for the centuries to come.

I totally see how it goes against self-preservation and how it's basically religion, putting that greater cause before the self, but I also think it's part of what makes us human and has allowed us to build civilizations in the image of shared gods.

[Adrià 🌲🌳 @AdrMmp]

Yes it was just an example, and yes more focused on western culture.
But I think we are in an inflection point in history (this century) where the values of the culture that builds strong AI and takes most solar system resources will dictate how most future humans (including ourselves) will live. Western culture is a good candidate to win (not guaranteed though), and although not perfect, has some norms and values that stop some of those brutal human impulses from happening, as you noted in your second case of loosing your freedom from killing your child.

[Adrià 🌲🌳 @AdrMmp]

@RokoMijic vindictive*

[Adrià 🌲🌳 @AdrMmp]

@RokoMijic You can also frame it in another way that biases it towards blue: the red vindicative dictator vs the blue benevolent ruler:

https://t.co/RQyoQ94U6f

[Richard Ngo @RichardMCNgo]

@HiFromMichaelV @catehall The scenario seems underspecified to determine the Schelling points.

Imagine the buttons are votes for a ruler. The red ruler is vindictive: if he wins he kills anyone who didn’t vote for him. The blue ruler doesn’t. So blue is the Schelling point.

But in Roko’s version red is.