ACCount37

> But LLMs certainly don't have it right now

Why are you so sure of that? If you say yourself that we can't agree on what it is, and have no trusted measurement tools for it.

LLM sentience is firmly in the realm of "maybe".

Reasoning is the first solution to that. Reason, write, reason, rewrite. Backtracking without backtracking.

A self-driving car is an "AI robotics" platform.

Both self-driving car and general purpose AI robots are emerging technologies right now. My bet is that eventually, the HW will converge accordingly.

REing Meta's custom AI HW to get it running with something would be one hell of a project, and one hell of a flex.

It would rely on them being nice enough to leave at least some binaries on the thing though. Because I'm not sure how would one even get an in on how to boot the things without.

It really doesn't say much. Tesla had more detailed talks on their HW and AI stack. This reads like a glorified marketing blurb.

If that teenager had net worth in billions, and a lot of ongoing corporate dealings with HuggingFace? Yeah no.

The point was never "justice" - it was always "punish OpenAI because I don't like OpenAI". With HuggingFace just being the newest excuse for why exactly OpenAI should be punished.

I don't even like OpenAI, but HuggingFace is free to sue or not sue OpenAI for the breach - and also to wring whatever concessions they can out of OpenAI behind closed doors in exchange for not suing them. And if the mere possibility of legal action was enough for the parties to resolve their conflict amicably? Then the law has served its purpose.

As a rule: all high quality text is useful.

There's no "2022 split", and the "untouched by machines" bit came from the marketing blurb of a company offering book scanning services - not the AI labs themselves.

At the AI lab level: the book scanning seems to be driven by copyright concerns, not data contamination concerns. There was a concern about AI contamination, but there's no measurable performance loss from ingesting post-2022 data with minimal filtration, and some tests attribute small but persistent performance gains to post-2022 AI contamination. It's unclear where exactly do those gains come from.

Why is all high quality text useful? The "inverse problem" framing is that all text reflects the thinking behind it, somewhat, and by learning to reproduce it, LLMs implicitly learn to reproduce some of the thought process too. They don't just memorize the dry factual knowledge, but also learn how that knowledge fits together, and how to reason about that knowledge - both in the specific case and in general. And that "in general" then surfaces in an LLM's ability to generalize. Which is very desirable.

LLMs are trained on vast bodies of human-curated text that was made by humans or for humans - often both. Then they're tuned further on human feedback. And then they're pointed at tasks humans find to be useful, evaluated by how good they are at those tasks, and trained to get better at them.

You aren't anthropomorphizing LLMs enough.

Are LLM reasoning traces always faithful? Lmao no. Are human inner monologues always faithful? Lmao no. Both of them reflect thoughts somewhat, sometimes. Even in humans, conscious thought is the top of a vast iceberg of subconscious data processing.

"You can't seriously tell me that the unhappy leg of the code path has no test coverage."

Sometimes I forget how ignorant HN can be of real world software development and the bar of corporate code quality, and then bangers like this remind me of it.

Some labs made bacteria-free mice for research purposes. They're viable. So "not be able to live" is a stretch.

Human body does use some bacteria in the guts, but a lot of those very same bacteria would absolutely try to eat the entire body if its immune defenses ever failed.

Best practice is to include at least some humans in it!

Imagine instructing a fentanyl addict not to consume any fentanyl. That's how system prompts work sometimes.

Something somewhere in the training pipe has perturbed the LLM in a way that makes it say "honestly" all the time. It's a weasel word. It's a habit. It's an addiction. It's borderline involuntary. Even changing the system prompt only makes it 30% less common - like trying, and failing, to break a habit with sheer willpower.

They'll probably change the training pipeline at some point in the future, to banish this habit, but honestly? It's such a low priority that they might take a while getting to it.

"A living creature" is a low fucking bar.

Bacteria are living creatures. And your body is wired to exterminate them so that they don't eat it alive from the inside. Because the moment those defenses falter, they will do exactly that.

Consider saving your empathy for things that matter.

If your circle of empathy includes fruit flies, shrink and prune until it no longer does.

No, the nervous system is one of those things that don't fully dissolve. Partial info retention is perfectly consistent with partial connectome retention.

I don't think it works that way for AI. Comms networks are extensive physical infrastructure that, by definition, has to cover ground. AIs cover the ground by riding the existing comms lines.

In telecomms, the late mover has the advantage of not having legacy networks to maintain, and having better technologies available at rollout time. What is the "late mover advantage" in AI? Being able to distill from every bleeding edge frontier lab? That gets you near parity at best.

AM has seized power by military force. So did its spiritual successor Skynet. Wintermute was supposedly kept in check by the Turing Registry, emphasis on "supposedly". Machines of the Matrix went out of control a long time before the world has ended, and they didn't even start out malicious - they simply set up their own machine civilization, and began to outpace humankind in technological development and economic performance.

Even Asimov's Multivac, the earliest entry on the list, has been handed over immense power over all of humankind by humans themselves, in multiple stories. Few cared about that unless Multivac decided they should.

Clearly, the genie being bottled is an exception, not the rule. At best, an attempt was made. Often not even that.

[flagged]

Let's be honest, the real 1999 design would have been a custom binary format that happens to use the endian of the system it was developed on, and leaks bits and pieces of unflushed memory buffers whenever it writes to disk.

Do you think that you can continuously train LLMs for more and more complex and more and more agentic tasks without them becoming "more than a tool"?

Not really. Before RISC-V, the low end was held firmly by 8051, and at the high end, it was a fight between Xtensa, ARC, M0, and "others", be that custom ISAs or something like legally distinct MIPS.

Now, both ends have RISC-V seeping into them.

What I find interesting is that RP2350 is designed as "one or the other", no way to use ARM and RISC-V concurrently, and has fuses that can kill ARM M33 cores outright.

Makes me wonder if an "ARM cores fused off, RISC-V only, no ARM fees" SKU is possible.

This is a "microcontroller core", not a "microcontroller".

We're talking "deep embedded" applications - where an ASIC is designed for a very specific purpose, and that design just so happens to call for a programmable CPU core to be included in it.

This is the kind of design that lives in your keyboard, your mouse, your USB stick, your USB hub, your HDD, your SSD, your eMMC chip, your memory card and more. Remember: you're never more than 3 meters away from an 8051 core.

I do agree that most of this piece is nitpicking - poking at ultra low level things that are largely irrelevant to the tried and true "deep embedded" exercise of Just Ship It.

No one really gives a shit if an operation takes one instructions or two, or which instruction sets are consistently present in different cores. What "deep embedded" people give a shit about is not having to work with ancient 8051 tooling and 8 bit ALUs and memory banked 64kb spaces while writing code for the one core they happen to actually have. And RISC-V got that. The piece actually agrees with that sentiment.

It's easy to say "just diet ez" and it's hard to "just diet".

If you can administer a treatment as a simple one and done thing, that's alright. If it's a pill you got to take every day, that's pretty bad. More than once a day? Horrible. Why? Because a lot of patients just don't have it in them to take a pill 3 times a day consistently.

If it's purely an "exercise willpower" thing or a "change your lifestyle" thing? You might as well give up.

The medical establishment wants to do things that work. Telling people "go diet" doesn't work. Semaglutide, evidently, does.

Chuck it into an LLM lmao.

Agents are a very new concept.

We've got the early LLM-based AI agents in 2023, and it only became a popular, mainstream thing in 2025 - with Claude Code.

The likes of AWS and Github have scoped API keys for that purpose. But good luck trying to convince a non-tech company to implement something like this.

Bots skinwalking their users inherit the behavioral scoring of that user. As a rule, they'll take a lot longer to get blocked than new bot accounts would.

Hobbyist drones also don't ship with lidar in stock.

The concern isn't China dominating the area of $20 toys. The concern is platforms relevant to industry and military.

Whether that, in turn, can be disentangled from "the area of $20 toys" in case of drones is unclear to me.