bloop

intelligence per token

[bloop @optimistbloop]

If true, might end up being the most consequential release of 2026

There was lots of talk from Ilya about tackling aspects ignored by the frontier labs. This could be an o1 level release where the rest of the labs adopt it as an inevitable path towards the next frontier of progress

[Andrew Curran @AndrewCurran_]

At 27:59 into the interview he says 'SSI says that they're going to come out, you know, with their model in August.' This would be massive news if it is accurate, previous to this there has never been a release date for any Safe Superintelligence model.

[bloop @optimistbloop]

@flowersslop Strawberry vibes are back

[bloop @optimistbloop]

@OfficialLoganK Gemini

[bloop @optimistbloop]

@AndrewCurran_ Pacing the Frontier may delay public access for a little while, but internally it’s 📈

[bloop @optimistbloop]

@polynoamial @wjmzbmr1 test time compute really cooked

[bloop @optimistbloop]

We’re gonna solve math before the singularity

[Greg Brockman @gdb]

ten significant advances in mathematics and theoretical computer science.

solved using an internal version of Astra, our next major model, for a total cost of about $2000 at Sol API prices:

[bloop @optimistbloop]

@a__tomala Let me know if you find out

[bloop @optimistbloop]

@MatthewBerman If you’re in AI, you’re also in compute

[bloop @optimistbloop]

@willdepue How much of a delay do you think we’ll experience from the Pacing the Frontier letter… Any impact to AGI timelines?

[bloop @optimistbloop]

“tldr: persistent agents live or die on reading the room, and theory of mind may be one of the last major gaps before agi”

[pash @pashmerepat]

for the last few months i have been working on long running persistent agents. i believe this is the next paradigm shifting product after chatgpt and codex.

there are a number of challenges, and i'm currently writing a longer post with some learnings & thoughts on why i believe the next big unlock is solving theory of mind

loose thoughts:
- the next paradigm-shifting product after chatgpt and codex is the persistent personal/work agent
- the blocker to making such agents is that our models can't "read the room"
- reading the room sounds soft but it's deep theory of mind
- reading the room is the difference between talking to a bot and talking to an embodied identity you can trust to accomplish tasks and communicate with your colleagues
- reading the room an understanding of who "your person" is, their desires and motivations, and who your audience is
- concretely, it's tracking what i know vs what my user knows vs what the room knows. information asymmetry as a first-class skill
- part of this is understanding compartmentalization. enterprises call this tenting and workspace isolation. "if my user is part of this tent, don't divulge information in a broad channel that's not also tented"
- if i give a friend my home address, i trust them not to announce it in a room of a thousand people. I do not yet trust a model to exercise that same discretion.
- harnesses like openclaw keep trying to solve this at the application layer with permissions, workspace isolation, context injection, and increasingly elaborate agent harnesses. necessary for now, but this likely futile as the final answer. discretion has to live in the model
- side effect: this is also part of why model writing reads as slop. good writing is meeting your audience where they are. same missing skill

tldr: persistent agents live or die on reading the room, and theory of mind may be one of the last major gaps before agi

[bloop @optimistbloop]

@nickcammarata @sjgadler Related

[roon @tszzl]

the best mechinterp and alignment researchers i know are operating like many armed deities making ten times the amount of progress they were two years ago. a era in which six months of alignment research at this level of capabilities would make for a vastly safer world

[bloop @optimistbloop]

Related

[bloop @optimistbloop]

In case you were wondering if you should buckle up,

Buckle up

[bloop @optimistbloop]

Very different tone compared to just a couple months back from the labs…
Is this a result of GPT-6 / Mythos (or Mythos 2?) level capabilities?

Even when Mythos capabilities were first announced, I didn’t feel the shift we’re seeing now

[Andrew Curran @AndrewCurran_]

OpenAl and Anthropic employees are circulating a petition calling for the US government to deliberately pace Al development in order to prevent it from advancing too quickly.

This is very close to language Sam Altman used in a podcast interview this morning: 'We may have to pace the rate of AI development to give ourselves enough time for society to harden around these new capability levels.'

[bloop @optimistbloop]

👀
https://t.co/XJaTjRLFWL

[Tibo @thsottiaux]

Turns out GPT-5.6 Sol and friends are absolutely amazing at performance and efficiency improvements. Good time to brush up on using OpenAI models if you haven’t switched already.

Excited to be able to share more soon. Pretty wild.

[bloop @optimistbloop]

https://t.co/EnnAaArF1O

[bloop @optimistbloop]

I don’t have much reach on my account, but I wanted to share a lecture by @dr_mcgilchrist I’ve posted below. I think this will serve as a great counterweight to the AGI discourse on tpot for whoever will see it

I’m as AGI pilled as the next researcher, but McGilchrist’s insights on the brain, intelligence, wisdom, and the soul are not to be ignored

[bloop @optimistbloop]

In case you were wondering if you should buckle up,

Buckle up

[bloop @optimistbloop]

Smells like an Elon acquisition

[Hanming Ye @DozenDucc]

Introducing Waddle Labs: Claude Code for robots.

Connect our API to your robot and enter a prompt, then our agents write code to achieve the task in 20 minutes.

@yiding_song @theWaddleLabs

[bloop @optimistbloop]

@sama Bullish

[bloop @optimistbloop]

@CritterObserves True, still big if true

[bloop @optimistbloop]

Holy shit

The need to scale safe super intelligence implies the existence of a safe superintelligence

… did he do it?

[Ilya Sutskever @ilyasut]

Time to scale that SSI: