HN user
I've seen some stuff that you wouldn't believe
The article is clear that LG is talking about banning all proxies, even voluntarily installed ones. A smart TV is a computer. Why should the owner of a computer not be allowed to run any program he wants?
It is, though. Why should I not be able to agree, with sufficiently informed consent, to route traffic through a connection I operate via a computer I own using software I've chosen of my own free will to install?
What's the difference between running a Tor exit node and a "residential proxy" except optics?
Look: I'm 100% for banning secret proxies that hide from the user, but stopping people knowingly and voluntarily running these proxies is a violation of fundamental software freedom tenets.
Is the mask ROM really going to be worth it over Carmack's high-bandwidth-flash concept? I mean, sure, I could be convinced I guess, but it's not obvious.
I love GNU Parallel too. I love how it's free software and how I can patch out the obnoxious citation notice.
Interesting. Is the speedup from specializing for the shape of Llama 3.1 or are they (contra my mental model) actually winning on burning in the weights?
You can pipeline in software too, and even with a dedicated circuit, you need to put weights somewhere, because you're not going to have a special partially-applied FP8 FMA unit for each of your trillion model weights.
I can accept the idea of specializing a circuit for a specific model shape, but I'm not seeing a need to specialize a circuit for the weights inside the shape.
Ah, right. I misremembered. Does Taalas have a meaningful advantage here then?
Cerebrus/Talas
They are fast, but they're still programmable accelerators, not a model burned into the gates.
What precisely is an ASIC supposed to do that a more programmable accelerator can't? Memory latency is memory latency: doesn't matter whether it's embedded in a cache-line wait state or some flipflop state machine. ROM isn't going to be faster than RAM either.
Likewise, for compute, is the ASIC somehow going to beat a systolic array? You can't have one circuit per weight: the die area and electrical fan-out would be insane. I'm not seeing how an ASIC specialized for a specific model would actually help much.
I mean, sure, we can build more specific accelerators, e.g. for softmax, but these work fine in the context of a programmable pipeline.
Yes, there are more exotic things out there, like optical matrix multiplication systems. Those are different. But above, aren't you talking about just doing conventional digital linear algebra, but with a model-specific set of circuits?
Perhaps not this for this parameter, but for others? Client side.
Codex is open source. You can build it yourself. You don't have to use upstream's parameter choices, compaction strategies, subagent heuristics, or whatever. Strange how people act like tweaks to these parameters are mandates when they're really just suggestions relayed to you via git.
For example: upstream recently changed the effort level hotkeys so M-. would stop at xhigh, not max. If you want max, upstream, you go to the /model menu. I didn't like this change, so I undid it locally. Easy peasy.
How do you know the "social value" is real value if you ignore the market?
Keep in mind that Venus is also an Earth-like planet with an atmosphere in the habitable zone of a sun-like star.
relentless norm of frankly insane positivity, not the ability for people to debate.
They are aspects of the same thing.
Debate implies that things aren't already perfect. "Why would positive people argue", a Californoid might ask, "instead of working together to make something great?".
My chief problem with California tech is that it's structured to reward schmoozing at the expense of ideas themselves and marketed as the opposite. It's unreality.
Debate in private, come to a decision and move forward.
This norm is terrible. It's a chief reason California tech companies become dysfunctional. Debate spreads common knowledge and ensures ideas get tested before execution. Hiding discussions is selfish, antisocial behavior that prioritizes your social image over idea quality and everyone's benefit. If you're working for someone who can't tolerate debate, leave.
It's weak leadership indeed that feels like debate undermines authority. A recipe for hell is
1. eradicating formal levels and hierarchy so everyone is "equal" and "ideas win", then
2. establishing a norm that debate never occur in public so that "ideas win" doesn't actually threaten whatever tyranny-of-structurelessness primate power dynamics emerge.
Thoroughly poisonous culture. A strong leader welcomes debate because his authority is secure.
It's good advice to "Prefer strict X in Y" for almost any value of X and Y. Lax DWIM stuff always comes back and bites you in the end.
I wish it were so. It's not.
For an object-level example, look at the difficulties the Russians had reconstituting their industrial base in the post-apocalyptic wasteland they found themselves after the 90s. It took them a decade to figure out how to restart production of the Tu-160M bomber, and they had all the original blueprints! Likewise, for a while, we forgot how to make this "FOGBANK" substance that's somehow important for nuclear weapons.
It's all too easy to forget how to do something if you stop doing it for a while.
Over time, the latter becomes the former.
TBC, the article isn't about "trust[ing] the database", but about the choice of TCB. There's nothing special about a "database" that makes it more vulnerable than any other API platform.
You're worried about SQL injection attacks? Why are you allowing untrusted clients to do anything other than call stored procedures? They can attack the DB? Why can't they attack the backend?
And sure: cryptographically binding a password to a username in storing a durable credential doesn't hurt, I guess, and it's cheap enough. (But why are you using bearer tokens in the first place?)
My objection is only the tacit assumption in the article that a "database" is some kind of distinct object that somehow accrues special superpowers and vulnerabilities relative to any other kind of service.
A database engine is just one choice of service implementation technology, like Go or Python. A database is just a service. Any service can be insecure.
Even if you're one of those people who believes the highest levels of performance require manual memory management (I'd disagree, but whatever), you'd do well to consider the performance of real-world GC against performance of the Arc/Box/dyn/clone-heavy "high-level Rust" [1] [2] that people and LLMs usually write.
Maybe you can write a carefully-tuned arena-and-slotmap that beats the pants off any GC. GC doesn't have to compete against P99.9 systems programming excellence. It has to compete against an endless soup of Arc, Box, and clone, and it will win this match without breaking a sweat.
I'm using malloc as synecdoche for all forms of allocation, including Arc<Box<T>> stuff.
Rust people: it's wrong to use a language that lacks memory safety and modern developer affordances. Do you really want to spend time debugging crashes and foist security problems on your users?
Me: You're right. Java has come a long way. Let's download...
Rust: No! No no on. Not like that!
---
Memory safety is a worthwhile goal, but combining it with manual memory management is wrong for most tasks. Just use a damn GC. Rust's safety-plus-malloc niche should be much smaller than it is.
ramming through deeply unpopular legislation at a breakneck pace for no apparent reason, lately
It's only a matter of time before EU-skeptical opposition parties achieve absolute majorities in critical EU constituency states. They're aware of this fact and are trying to adopt as much of their agenda as possible in the time they have remaining.
I don't accept that social media is self-evidently bad. Not everyone lives inside the bubble in which its wrongness is obvious.
You can't ascertain what's true without free speech.
Suppose you start with a true belief. This belief, like any other, has to propagate from person to person to sustain itself. The truth is messy. During each hop, people present the best, cleanest version of the truth in the belief. After enough hops, the belief stops being the best version of the truth and starts being a truth-flavored falsehood.
If you want to optimize for your beliefs being true, you can't just prohibit their opposite. Even if you don't care about truth per se and want only for your beliefs to spread fastest, you should want to make sure your beliefs stay true-ish, because if you decouple your beliefs from truth, then they become so mutated and useless that they prompt people to whom you propagate the idea to seek out other ideas on their own.
Yes, it's easier than ever for bad ideas to spread. What people who use words like "disinformation" miss is that this ease cuts both ways. What makes it easy for opposite beliefs to spread makes it easier for false mutations of their own beliefs to spread.
Even if you can ban opposite beliefs, you can't ban mutated forms of your own beliefs: attacking the mutants looks like attacking the truth forms, yet, because the mutations smooth over the messy parts of the truth, the mutant versions of your belief will out-compete the original within the bubble of your ban.
Truth alone doesn't stop the mutants because it operates on long time horizons and small gradients. Only strong contrast can distinguish truth from appealing falsehood, and only competition with opposite beliefs can establish it. Trying to establish this contrast within your belief's framework against a near-mutant looks like gatekeeping, pedantry, or disloyalty. The near-mutant is always too close to its parent to justify the social cost of an all-out fight.
Furthermore, once a mutant version of your belief has taken over, the original version itself looks like heresy. Deformation occurs in small steps because it happens unwittingly in response to intuitive incentives. Reformation occurs in large steps because it happens wittingly in response to observing incoherence. You can't ban near-mutants, but you can ban far-mutants, and reformers trying to jump from the endpoint of a long line of near-mutations back to the original form get the original form and themselves banned.
So now what? If you think it's hard to defend your true beliefs against opposite beliefs, think how much harder the job will be once you can't wield truth as defense. By default, beliefs win and lose as they become extractive and appealing in cycles. Truth allows an idea to win despite being costly. Without truth as a benchmark, why would anyone prefer your belief over upstarts that promise fewer rules and more fun?
If anything, people will prefer opposite beliefs to yours because you've been the one calling false beliefs true just because they descended from true beliefs and because you're the one telling them to shut up and get in line. Even if you start out with the true belief and the opposite beliefs are all false, you lose.
People, even in tech, even in AI, are allowed to have opinions that differ from yours. Cancellation doesn't work anymore: you can't get away with presenting your side as normal and other side as deviant and let ostracism win your idea's battles without your idea having to fight for itself.
The reason it's cringe is that you can really only understand an idea by placing it in tension with other ideas. Remove your idea's competition and you remove an incentive to explore the weaknesses of your own. Ideological protectionism, just like the economic kind, breeds weakness. Your idea mutates and, without fitness feedback, drifts into a ridiculous parody of itself. Your idea ceases to be a thing that can live on its own and comes to depend on the protectionism for it's survival. Yet, the more ridiculous your idea becomes, the more protectionism it needs to compensate. One day, your idea collides with other ideas (which are still out there) despite your best efforts to shield it attack, and when it does, you're shocked by how weak your coddled, mutated idea really is compared to the original form one might remember.
All the people out there criticizing Grok, or Grokipedia, or whatever for espousing the wrong ideas are ultimately undermining their own. Even if you don't believe in high-minded mumbo-jumbo about the value of free speech, even if you just want your side to win, trying to shame models like Grok into not existing is foolish and undermines your goals.
From a computer science POV, it's spooky how Turing-completeness arises spontaneously out of the most mundane data-processing machines. You look at UTS#35, see "Transforms provide a set of rules for transforming text via a specialized set of context-sensitive matching rules." and think, "Ah! Rewrite rules! Those are often Turing-complete! Maybe this one is too!".
And so it is (https://en.wikipedia.org/wiki/Semi-Thue_system#Undecidabilit...).
It's a scary and wonderful part of our world that you can bootstrap so much complexity out of a little iterated self-reference.
My only quibble with the article is in this line...
The surprise is that it lives in a data format for locale files, shipped in every OS, whose specification doesn't mention the possibility.
... I'm not surprised. After all, the processor that interprets the data format is Turing-complete not only in its instructions, but in the page table! See https://github.com/jbangert/trapcc
If anything, when you build a system and it starts to get complex, you have to go out of your way to ensure it's decidable and can't accidentally bootstrap the universe.
GPT-Live-1 is the first version of a new generation of models, and we believe the full-duplex architecture + delegation enables entirely new ways of human-AI interaction.
Awesome. Are you guys able to share anything about the model architecture? I've been interested lately in split-transformer RVQ-based conversational agents, e.g. via stuff like https://arxiv.org/abs/2412.10208 (ResGen) and https://arxiv.org/abs/2603.18090 (MOSS-ITT) and of course Moshi (https://arxiv.org/pdf/2410.00037).
Intuitively, decoupling semantic and audio-timeslice-space generations with coupled but distinct histories is right model architecture, not just for these sorts of assistants, but for domains like robotics too.
This is the opposite direction AI should be going.
There is no moral obligation, in any domain, to refuse to make a product that adults, with full informed consent, find useful and purchase. Who are you to say you know better than the market?
Intellectual and economic accomplishment aren't exactly parallel vectors, but you can't expect me to believe they're orthogonal. A mediocre tech CEO is going to run intellectual circles around some random guy you pick out of the DMV ticket queue.
And, yes, some people are smarter than others. Some people are a lot smarter than others. Also, smart people are rare. Very smart people are very rare. These are basic facts of life that continue to exist whether or not you believe in them. Ignore them at your own risk.