> without interoperability whats the point

Competition lowers prices and increases value.

A monopoly or design-by-committee would just ossify.

I can't hook up my DSL modem to Comcast or Community Fiber but that's fine.

Install Brave and turn off the extras in the Preferences?

It takes two minutes and doesn't bother you again.

Spend an extra two minutes to set up a sync chain, though - it's secure and handy.

Half the news stories I see embed a Tweet and roughly none embed a Thread (what's the correct singular?).

Maybe it's relevant outside of news?

Anything to sabotage the peace talk summit on Friday.

A BOLO for false flags was issued two weeks ago but this is pretty weak.

Even if it were true you'd hold this for a week unless you're a war whore. But NYT seems to worship death and the profits that come along with it.

Are you assuming his claim is true? He didn't get rich telling the truth. Three different AI apps have reached #1 this year. Deepseek, Perplexity, and GPT.

I know this is an alien concept to most people here, but it would be nice if people would actually, you know, read the papers first? I know nobody does this, but, could people at least try?

First off, this isn't peer reviewed. So it's not "actual, careful research", it's "not yet analyzed to determine whether it's decent research".

Secondly, despite what they call it, they're not dealing with LLMs at all. They're dealing with Transformers, but in no way does it have anything to do with "language", unless you think language is repeated mathematical transforms on random letters.

It also has nothing to do with "large". Their model that most of the paper is based on is minuscule, with 4 layers, 32 hidden dimensions, and 4 attention heads. A typical large frontier LLM has maybe 128 layers, >10k hidden dimensions, and upwards of 100 or so attention heads.

So right off the bat, this has nothing to do with "large language models". It is a test on a toy version of the underlying tech.

Let us continue: "During the inference time, we set the temperature to 1e-5." This is a bizarrely low temperature for a LLM. Might as well set it to zero. I wonder if they have a justification for this? I don't see it in the paper. Temperatures this low tend to show no creativity and get stuck in loops, at least with "normal" LLMs.

They train it with 456976 samples, which is.... not a lot. Memorization is learned quickly in LLMs, while generalization is learned very slowly (see e.g. papers on "grokking").

Now here's what they're actually doing. They have two types of symbol transformations: rotation (for example, ROT("APPLE", 1) = "BQQMF") and cyclic shifts (for example, CYC("APPLE", 1) = "EAPPL".

For the in-domain tests, they'll say train on ROT, and test with ROT. It scores 100% on these. It scores near-zero on the others:

Composition (CMP): They train on a mix of two-step tasks: ROT followed by ROT; ROT followed by CYC; and CYC followed by ROT. They then test with CYC followed by CYC. They believe that the model should have figured out what CYC is doing on its own and be able to apply CYC twice on its own.

Partial Out-of-Distribution (POOD): They train on simply ROT followed by ROT. They then task it to perform ROT followed by CYC. To repeat: it was never traiend to do CYC.

Out-of-Distribution (OOD): They train simply on ROT followed by ROT They then task it to do CYC followed by CYC. Once again, it was never trained to do CYC.

The latter two seem like grossly unfair tests. Basically, they want this tiny toy model with a "brain" smaller than a dust mite's to zero-shot an example it's had no training on just by seeing one example in its prompt. That's just not going to happen, and it's stupid to think it's going to happen.

Re, their CMP example: the easiest way for the (minuscule) model to learn it isn't to try to deduce what ROT and CYC mean individually; it's to learn what ROT-ROT does, what ROT-CYC does, and what CYC-ROT does. It doesn't have the "brainpower", not was it trained to, to "mull over" these problems (nor does it have any preexisting knowledge about what a "rotation" or a "cycle" is); it's just learning: problem 1 takes 2 parameters and I need to do an an offset based on the sum of these two parameters. Problem 2... etc.

The paper draws way too strong of conclusions from its premise. They do zero attempt to actually insert any probes in their model to see what their model is actually doing (ala Anthropic). And it's a Karen's Rule violation (making strong assertions about model performance vs. humans without actually running any human controls).

The ability to zero-shot is not some innate behavior; it is a learned behavior. Actual LLMs can readily zero-shot these problems. And by contrast, a human baby who has never been exposed to concepts like cyclic or rotational transformation of symbols could not. One of the classic hard problems is how to communicate with an alien intellect - if we got a message from aliens, how could we understand it? If we wanted to send one to them, how could we get them to understand it? Zero-shotting communicative intent requires a common frame of reference to build off of.

If anyone tells you our Country is at risk because people in charge of mass government censorship campaigns got fired then they're your enemy domestic.

Unless you're one of them and they're your partner in crime.

We must move past not trusting spooks to ignoring spooks, except as to understand the situation vis-a-vis enemy intelligence.

PS Free Snowden!

Just yesterday there was a news story about
Predditors organizing to mass copy a YouTuber's content to try to wreck his revenue.

A court forced Reddit to hand over their identification and he is suing them. Ethan somebody.

There's quite an ethos over there about organizing crime and apparently if it's leftwing they just leave it alone. e.g. https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.reddit.com%2Fr%2Flgbt%2F... nobody pushing back against crime.

Most neutrally the company may not want to deal with subpoenas and they don't think the crimes are justly constructed.

But it's weird which kinds of illegal-in-some-jurisdictions subs get adult gated and which ones are freeworld. e.g. drugs are legal in some countries and being gay is illegal in others. Region settings seems to have no effect. Rightwingers get ejected post-haste. Tech subs are often great, sometimes conquered by schizo mods. So unpredictable over there. I go there sometimes like I go to failed neighborhoods sometimes - only when the need arises.

The fatigue seems to be mostly about other humans who believe LLM's are omniscient and inflatable.

Or even good conversationalists.

As an information retrieval tool or a radiology diagnosis assistant, sure, hardly anybody is complaining.

OK, maybe some lesser radiologists.

Unless they're holding off because of IP concerns, that doesn't make any sense to me. If the tools work well enough to be worth using on personal projects, why not use it on paid work?

This is my point. Don't use it if it wastes time or produces bad results, use it when/where it saves time. One easy way to do this is to copy your source repository and tell the coding agent to go write test cases or implement a feature or whatever and keep working on it until the code builds and passes, and while it's working you do the work yourself. When it's done (it will almost certainly finish before you), git diff the result and decide whether to use what it did or what you've done 20% of. The time investment for this is negligible. Or, what I tend to do is to set the LLM working while I catch up on email, write design docs, attend meetings, etc.

Hype or lack thereof is irrelevant. If makes you more productive, use it. If it doesn't, figure out why not. If it's because you should be using it differently, do that. If it's because the tool just sucks try a different one, or ignore it for a few months until it gets better.

It is for me. Why wouldn't it be for you? If the LLM produces code that doesn't meet my rather picky standards, I tell the LLM to fix it, or I fix it. Either way, the code is going to be basically indistinguishable from what I'd write before I send it for review. I guess it's possible that the LLM could write better code than I would write, but it definitely can't be worse than what I would write, because I won't allow that.

I've been a professional software engineer for 35 years, and been involved in hiring for all but the first two, in several different companies from tiny startups to giant corps (IBM and now Google). Maybe computer engineering is different (I doubt it), but in all that time I've never seen or heard of a company that cared about GPA, because it's a really lousy predictor of ability. Sometimes recruiters use GPA as a screening tool when they have absolutely nothing else to go on, but that's the only use of GPA I've seen.

Companies mostly want to see if you can really do the job, and the best evidence is showing that you have done the job, i.e. you have real, professional experience as a software engineer. Failing that, they want to see evidence of you having built stuff, so a Github portfolio (or, before that, a Sourceforge portfolio). The best way to get said professional experience is to do an internship or two while you're still in school. Spend your summers working (and getting paid!), so that when you graduate you have that experience to point to.

Yes, absolutely, to the internship. Meh to the GPA.

That said, I'm not surprised it's gotten a little tougher. AI tools can't replace experienced SWEs yet, but under the supervision of an experienced SWE they definitely can do most of what interns and entry-level SWEs usually do.

I don't understand what this means. How do you "dabble" and what would it mean to "pivot"?

I use AI for stuff it does well, and don't use it for stuff it doesn't do well. Knowing which is which requires investing a little time, but not much.

Why should any of that matter? If using it makes you more productive, do it. If it doesn't, don't. This isn't like a new language where using it requires a significant commitment because you'll have to maintain it in the future. The code you write with AI should look basically the same as the code you'd write without it.

100% this. I'm a vegetarian, the sort of person they think should be buying their products, but their products disgust me, because they remind me of meat. I don't want to be reminded of an animal corpse while I'm trying to enjoy a tasty meal. Why mimic the thing I don't want to eat?

(I'll only speak re: vegetarians below, but I expect vegans are similar)

I would ask non-vegetarians: imagine that you live in a world where people ate toddler meat. Real human toddlers, slaughtered for their meat. The vast majority of people in your situation would of course avoid eating them. Some may be radical anti-toddler-meat campaigners. Others may silently accept that they're not going to change the rest of the world. Either way, you let people know that you don't eat toddler meat so they don't serve it to you. But hey, some friendly cooks feel bad that you don't get to enjoy toddler meat! So they make a baby meat substitute that looks and tastes exactly like toddler meat! They package it in packages with pictures of dead toddlers on it, but with labels "No toddler included!" And then they expectantly wait for you to thank them and praise them for finally making toddler meat that you can eat - rather than being disgusted by the whole concept and wanting some non-baby related food that doesn't make you think about dead toddlers while you eat.

That's not the situation *all* vegetarians are in, but it is the situation that a *lot*, dare I say most, vegetarians are in.

I think a lot of non-vegetarians cooking for vegetarians are just frankly confused about what to offer us, as they have trouble picturing of a meal without meat. It's really simple: you know how fatty, salty, carby, umami-rich stuff tastes really really good and leaves you feeling satiated? Yeah, just make something that's fatty, salty, carby, and umami-rich that doesn't involve meat, and your vegetarian friends will be happy ;) Like, pick anything carby, cook it with a tasty fat that pairs well with it, add salt and something umami-rich (things like mushrooms, nuts, tomatoes, yeast, nori, olives, nuts, cheese (vegetarians only), spices, etc etc), and viola, that's a good vegetarian dish ;)

  Of course, *to make it healthier*, and a more "adult" taste, you'll want to include non-carby veggies (which, per unit *dry mass*, are actually highly protein rich, with e.g. freeze-dried broccoli being well more protein rich than your average grade of ground beef without its water, and freeze-dried watercress being up there with fish - they're just heavily watered down). Veggies also add umami. You can also - optionally, but it's not at all a requirement - include high-protein things like tofu, tempeh, seitan, TVP, etc. But protein deficiency is not common among vegetarians or vegans in western society (the main risk is iron deficiency, particularly for vegans, and - exclusively for vegans - B12 deficiency, but only if they don't eat anything fortified with B12, though B12 fortification is common).

Am I the only person getting whiplash that we're rediscussing the exact same thing when this concept was already proposed as Breakthrough Starshot, and was big in the press at the time, incl. on this site?

Anyway, you still have to have the energy to transmit back, which was proposed to be from a RTG: My hot take: since you already have to have a (comparably) big sail anyway, which means booms to maintain its structure, use 232U to get many times the energy per gram as 238Pu (38,9 MeV vs. 5,6 MeV), at the cost of a hard gamma from 208Tl, and put it out on the ends of the booms, with appropriately radiation-hardened electronics. You could also do double-duty with an alpha sail (alpha emitter backed by a thin low-Z material to create net thrust from alpha emission)

Also, if the 232U is in the form of a compound that's sufficiently soft for fast diffusion at its equilibrium temperature, you can diffuse out the 220Rn and avoid 208Tl's hard gamma altogether. This costs you (you only capture 16,6 MeV of alphas), but not only does it avoid the hard gamma, but it also means that you don't retain the mass of the stable decay products, so your craft gets significantly lighter over time. (At one point I was considering urania aerogels to lose the radon instead of "soft" high-diffusion materials, but the data was suggesting that the aerogel would quite quickly self-pulverize and densify)

232U is readily producible (indeed, it's a waste product in thorium reactors); main issue is just that it's a pain to handle. But for something like this, you're dealing with microscopic amounts.

Also, it's questionable how many of the civilians were really civilians, given the Japanese "Ketsugo" plan, which organized all men 15-60 and all women 17-40 into combat units, trained millions of civilians, including children, to fight with bamboo spears, farm tools, explosives, molotov cocktails and other improvised weapons and had the slogan "100 million deaths with honor", meaning that they expected the entire population would fight to the death rather than surrender to invasion.

Would that actually have happened? No one knows. What is certain is that Allied knowledge of Operation Ketsugo caused war planners to increase their estimates of Allied casualties because they assumed they'd be fighting the entire population. Experience with Japanese soldiers who consistently fought to the death validated that the Japanese culture just might be capable of that.

That was an important part of the context in which the decision to drop the bombs was made.