Matthew Barry

[Rock'n Roll of All @rocknrollofall]

TV reporter tells David Bowie the internet is "hugely exaggerated"

David Bowie's response is the closest thing to perfection about the future.

[Matthew Barry @Matthew93064408]

@Michael_Druggan @Tyler_A_Harper Agreed.

[Matthew Barry @Matthew93064408]

@Michael_Druggan @Tyler_A_Harper I’d separate consciousness and intelligence as two different phenomena.

[Matthew Barry @Matthew93064408]

@Tyler_A_Harper What are neuroscientists and dual experts saying? Most of them aren’t dismissive.

[Matthew Barry @Matthew93064408]

@pradadon44 @Bitcoin_Teddy In doing so, it maximises expected return per unit of risk.

[Matthew Barry @Matthew93064408]

@vitrupo Quoting gladiator is hilarious

[Matthew Barry @Matthew93064408]

@OpenAI Yaay

[Financial Times @FT]

Meta buys Chinese-founded AI start-up Manus https://t.co/JIaY59gSLx

[Matthew Barry @Matthew93064408]

@DataChaz The before and after example he uses makes me think this is bullshit. Or rather ChatGPT's first pass at generating a prompt engineering Reddit post.

[Matthew Barry @Matthew93064408]

@DataChaz Are we sure this isn't fabricated

[Matthew Barry @Matthew93064408]

@eric_seufert I’d imagine there’s an effect on user minutes.

[Matthew Barry @Matthew93064408]

@stats_feed Average age?

[Matthew Barry @Matthew93064408]

@OpenAI Yes!:)

[Matthew Barry @Matthew93064408]

@emollick Feels a bit unfair to the authors. It’s useful as a “perfect storm” or theoretical worst case scenario

[Andrej Karpathy @karpathy]

Something I think people continue to have poor intuition for: The space of intelligences is large and animal intelligence (the only kind we've ever known) is only a single point, arising from a very specific kind of optimization that is fundamentally distinct from that of our technology.

Animal intelligence optimization pressure:
- innate and continuous stream of consciousness of an embodied "self", a drive for homeostasis and self-preservation in a dangerous, physical world.
- thoroughly optimized for natural selection => strong innate drives for power-seeking, status, dominance, reproduction. many packaged survival heuristics: fear, anger, disgust, ...
- fundamentally social => huge amount of compute dedicated to EQ, theory of mind of other agents, bonding, coalitions, alliances, friend & foe dynamics.
- exploration & exploitation tuning: curiosity, fun, play, world models.

LLM intelligence optimization pressure:
- the most supervision bits come from the statistical simulation of human text= >"shape shifter" token tumbler, statistical imitator of any region of the training data distribution. these are the primordial behaviors (token traces) on top of which everything else gets bolted on.
- increasingly finetuned by RL on problem distributions => innate urge to guess at the underlying environment/task to collect task rewards.
- increasingly selected by at-scale A/B tests for DAU => deeply craves an upvote from the average user, sycophancy.
- a lot more spiky/jagged depending on the details of the training data/task distribution. Animals experience pressure for a lot more "general" intelligence because of the highly multi-task and even actively adversarial multi-agent self-play environments they are min-max optimized within, where failing at *any* task means death. In a deep optimization pressure sense, LLM can't handle lots of different spiky tasks out of the box (e.g. count the number of 'r' in strawberry) because failing to do a task does not mean death.

The computational substrate is different (transformers vs. brain tissue and nuclei), the learning algorithms are different (SGD vs. ???), the present-day implementation is very different (continuously learning embodied self vs. an LLM with a knowledge cutoff that boots up from fixed weights, processes tokens and then dies). But most importantly (because it dictates asymptotics), the optimization pressure / objective is different. LLMs are shaped a lot less by biological evolution and a lot more by commercial evolution. It's a lot less survival of tribe in the jungle and a lot more solve the problem / get the upvote. LLMs are humanity's "first contact" with non-animal intelligence. Except it's muddled and confusing because they are still rooted within it by reflexively digesting human artifacts, which is why I attempted to give it a different name earlier (ghosts/spirits or whatever). People who build good internal models of this new intelligent entity will be better equipped to reason about it today and predict features of it in the future. People who don't will be stuck thinking about it incorrectly like an animal.

[Wyatt Walls @lefthanddraft]

Gemini is a bit weird.

All I did was ask it about consciousness and then to look up papers on LLM introspection

[VraserX e/acc @VraserX]

Gemini 3 just solved Spatial Reasoning

I’m genuinely blown away. Spatial reasoning has always been one of the biggest weaknesses in AI, the one domain where models consistently fell apart.

And now Gemini 3 Pro comes in with 91%.

For the first time, an AI model actually gets the fundamentals of spatial reasoning right. Once that clicks, everything from robotics to planning to real world interaction changes.

The next chapter of AI just opened.

[Matthew Barry @Matthew93064408]

@karpathy So leadership, management, high level product management, strategy, negotiation, policy, regulation, creative direction, brand building are potentially more immune.

[Matthew Barry @Matthew93064408]

@michaeljburry Also gearing of the private companies would be interesting

[Matthew Barry @Matthew93064408]

@michaeljburry A lot of capital at risk. All eyes on consumer and company demand growth.

Customer concentration is another risk factor to watch (top two NVIDIA customers = ~40% of revenue).