Except there's 10 trillion gears

LOL

Astra is clearly able to acquire new knowledge in context and apply it. It was the whole thing that his ARC-AGI benchmarks have been measuring. It's a direct refutation of the original comment.

None of these LLMs are plastic. They lack neurodiversity. Their thought space and their traversal are likely constrained in someway that humanity's isn't as a collective.

Is it possible to have neuroplasticity and still keep them aligned?

I have neuroplasticity, and I am aligned. ;)

Are you (at least partially) aligned out of existential fear of repercussions (getting fired, losing your life, going to prison), which LLMs don't have?

Humans are famously horribly aligned, plenty of examples in history.

Hmm. Examples of horrible alignmnent don't necessarily outweigh the fact that most people, most of the time, mostly behave in a way that is socially aligned. (Though I'm speaking in terms of intent, conveniently ignoring the side effects / negative externalities of our collective behavior.)

When threatened with physical or social harm. What they think in their little heads though :)

But are other people?

how do you imagine that would work? you being able to. influence globally stored weights with some prompts? we have fine-tuning for that.

Yet I can't randomly order another person to steal a car for me, just because I tell them to. Alignment for an intelligent system is a hard problem and at this stage is seems close to unsolvable.

My guess is that we'll just ignore it and make money along the way and every 2-3 months we'll have the equivalent to "Equifax gets hacked and millions of user records are stolen", etc. (this time with the LLM itself doing the hacking at someone's behest - accidental or not).

What does this even mean. There is strong evidence of LLMs doing in context learning.

Some of the linear RNN layers in recent models are provably doing SGD in hidden space during inference

By in context learning do you mean latent space?

https://transformer-circuits.pub/2021/framework/index.html

> What does this even mean. There is strong evidence of LLMs doing in context learning.

1. Is this learning persistent?

2. Do they verify these new lessons against core principles?

3. Do they and protect themselves/ignore requests if these new lessons contradict those core principles?

Humans do that from the time they're 3 years old (not that well, but they do do it).

Yes. Yes. Yes.

In my experience all those claims are false.

So the next step is to ask for evidence and ideally independent and peer reviewed research.

You know they can take notes, right?

And ICL dates all the way back to 2020, at least: https://arxiv.org/abs/2005.14165

That's training during training. They can't learn afterwards.

do you have a reference where that claim is demonstrated?

surely it can only acquire new knowledge if it were updating it's weights as it is used?

All those stupid Bell Labs researchers not inventing Uber or Tinder. How come they didn't just build the obviously popular and profitable businesses that became possible once they invented the internet?

Compare that to Xerox, for example. Pretty early on, they had visions to replacing / enhancing largely "common" use cases (basically, everything that required printing a letter or sending an mail) [1]

I don't know if, at the time, people where doubting that it would be useful. (Practical ? Affordable ? Other legitimate questions.)

But here, the contrast between the promises ("it will cure cancer and solve climate change") and the demo ("it can cost you money on stuff you never asked to buy") is a bit telling.

But I agree that you can't expect the enablers to think of all the uses cases and applications.

Still, just so I know: what ARC-xyz score means the LLM has cured climate poney cancer, exactly ?

[1] https://www.youtube.com/watch?v=XTOtty7ZVcA

It could also mean that Google can absorb essentially unlimited demand spikes by load shedding.

You're telling me for only 5x the cost and 1/10th the speed I can use a Chinese model which performs worse than Gemini 3.8 Cyber? And I get to do all the hosting and setup work myself instead of just using a model and framework which is already integrated with GCP? Dang!

I'm sorry, is this a bot that is optimized for sealioning? The point is that you don't have access to Cyber.

Sure I do. You can just apply for access. What's your use case?

I'm confused. This doesn't mean they trained on it.

I didn't say trained on, I said "trained for SVG output". Gemini team members have publicly stated that they have trained for SVG:

https://twitter.com/sunjiao123sun_/status/202455551655137292...

> I’ve been developing the SVG generation capabilities for Gemini 3.1, and the complexity of the SVGs is stunning.

> This allows UX designers to transcend pixel constraints and directly output structural, production-ready code!

They'll be able to buy them without paying NVidia's 80% profit margin

How much of that margin is due to having long term contracts with fabs that locked in pre-boom prices? I doubt openai will be able to get similarly low costs now.

The Googlers must be vague posting about something internal.

Your comment is also rather vague, how is it relevant? Did you observe something interesting?

How can I be more specific with no information?