beloch

Let's put gaming and all the other uses for compute that require low latency, etc. off the table and just talk about AI compute. People can and do run open models on their own hardware. Altman is, not unseriously, arguing that this hardware would be better utilized if it were concentrated in data centres rather than distributed into smaller, private clusters that might not reach full utlization.

Now, let's take privacy concerns off the table. OpenAI's complete lack of respect for privacy, authorship, copyright, etc. mean that you simply can't entrust private data to OpenAI. Full stop. Anyone with privacy concerns gets a free pass.

Finally, we're dealing just with people doing LLM computations on data that's not sensitive. Why aren't these people doing things in data centres? There are likely a variety of reasons, but one is likely price. If somebody does a lot of stuff with local models, they've probably found an application where a bigger closed model doesn't offer compelling advantages and the cost, purely in terms of filthy lucre, of running a local model is lower. These are the users Sam Altman can legitimately target. What Altman is arguing is that these people are paying too much.

How should he target them? Become the Steam of compute. Save them money and time. If Altman wants these people's RAM because he can use it more efficiently, he needs to offer them the same services they can get from their own hardware, cheaper, and with less trouble.

As one would hope, given that it shared the same time setting and some characters. It's truly a shame they never managed to get Patrick Stewart in for a guest appearance.

Some TV shows are now being designed to be watched as background while the viewers are pecking away on their phones. Plots are simplified and repeated. Viewers are often shown and then told important points multiple times, just to make sure they don't miss them.

Discovery did this some of the time. There were episodes where they would show, then tell, then tell again all within a few minutes. Not every episode was done this way. I suspect they were experimenting and it did not go well. Anybody who was actually giving the show their full attention was bound to be irritated. Discovery was a low point in modern Trek for multiple reasons. They initially decided to forgo an ensemble cast and bet it all on a protagonist driven narrative. This severely limited the diversity of viewpoints in the first couple seasons. The writing was bad for entirely avoidable reasons. People who watch trek pay attention to the universe's rules. You can't beam through shields, you need to have a reason for why the away team can't be trivially beamed out of dangerous situations, etc.. The people writing Discovery did not bother to do this on a repeating and continual basis. Again, it was insulting to anyone who wasn't pecking away on their phone. I wouldn't say the writers for Discovery merely lacked life experience. They clearly hadn't bothered to watch previous Trek shows.

Strange New Worlds is better written, but a little too reliant on high concept for their most recent season. Every episode seems to have a gimmick. Still the cast is good and I look forward to them wrapping up Pike's story and, presumably, setting up a reboot of TOS.

Lower Decks is the best Trek show in recent memory. Thoroughly entertaining from start to finish. I can't recommend that one enough, just in case anyone skipped it because it was animated. Intelligent. Inventive. Hilarious. The people who made Lower Decks really put a lot of love into it.

Edit: I forgot to mention Picard, which is unforgivable. Some of the seasons were a little uneven, but some were fantastic and it was truly a treat to see Patrick Stewart as Picard one last time.

A fingerprint encrypted USB drive is just a bad concept to begin with.

1. Anyone who has access to the drive likely also has a nice sample of your fingerprint on the same device.

2. Hardware encryption implementations are often unknown, flawed, or completely missing. You shouldn't place trust in obscurity.

3. Hardware encryption often adds an unnecessary point of failure. I've seen drives with hardware encryption implemented on a daughter board become unrecoverable because the daughter board failed, even though the drive itself was just fine.

Put a software encrypted volume on a normal storage device and call it a day. Don't pay extra to do worse.

TNG put writing first. They went after interesting stories and didn't try to force them into a formula for ratings. Sometimes there was fisticuffs and phaser fire, but many of the best TNG episodes didn't involve violence at all. It was an optimist's take on the future. The cast of TNG was great, but Patrick Stewart's acting chops let them do episodes that another show couldn't have pulled off. He changed what was possible.

The quality of writing for DS9 was also quite good but, after trying the TNG formula for a season or two, they broke with the writing-first mentality and adopted an action sci-fi formula. The showrunners feared they'd lose ratings if an episode didn't have at least some action. It worked for DS9, but later Trek series have had a tough time shaking free of the action sci-fi formula. The problem with formulas is that, while they produce individual episodes or shows that are exemplars of their form, the sameness eventually causes viewers to drift away.

Strange New Worlds has shown it's willing to do some pretty wild things (e.g. turning its cast into muppets, adding gore and horror elements, etc.), but they have not messed with the formula in a meaningful way.

If you're looking for something outside the Trek formula, one recommendation I'd share is "The Foundation".

"At most companies, only someone with decades of experience under their belt would be trusted with a project this size. Patel thinks that’s a problem.

“If you don’t hire the early-in-career [workers], and you just let your workforce continue the way that they were, where do fresh ideas get injected? How do you challenge the status quo?” he said."

-------------------

This would be an interesting approach for a bleeding edge startup that's betting big on producing something novel or going bankrupt. If you're going to bet the company on one guy and some AI, that's bold. He'd better have some serious chops and not be merely a slick salesman.

However, Cisco is an infrastructure company. Status quo is what most people want from an infrastructure company. You don't want your Cisco stuff to suddenly start doing new, innovative things that require a complete redux of everything built on top of it. This, of course, assumes that having one guy do a refactor of your entire code-base produces something "innovative" rather than a cavalcade of snafu's.

It's probably not the best time to be holding CSCO.

Silly billionaire, fair use is for people, not megacorps!

But seriously, fair use is for students, teachers, comedians, and all the people who can't pay or couldn't possibly pay enough for permission from copyright holders. A corporation training a closed model for profit can, and should, pay. If American copyright law is too onerous for megacorps to contend with, maybe it should be reformed.

Scale seems to be the real issue. Given that there are relatively few actors with the scale to cause serious problems (and those actors have the resources to shrug off minor penalties), perhaps more targeted legislation is necessary. e.g. Different penalties for petty scraping and grand theft scraping. Of course, this would necessarily target the people who pay for election campaigns, so it will never happen.

3. Some applications are going to work just fine with smaller models on local hardware. If people still have machines with the memory to run those models laying around, that will cost OpenAI business. Also, once RAM suppliers catch up and Moore's law resumes operation, the capability of local hardware is going to steadily improve. Were it not for the AI build-out, typical configurations for new PC's would likely have moved on from 32GB by now.

4. Everything is training data. If that data isn't going to OpenAI, they're not training on it.

Also, it's important to remember OpenAI's failure might actually be the best thing for everyone else in the AI field. e.g. Some new innovation/approach could come out of left field and eat OpenAI's lunch. They've bet the farm (and a lot of other peoples' farms) on a brute force approach. Altman needs your RAM because, if he doesn't control the brute force, the brute force approach might not work only for him.

"In order to make one of these things work with natural (unenriched) uranium, which was all that they had to work with at the time, they had to get a lot of it situated close together but interspersed with neutron moderators such as pure graphite or heavy water. The Germans just didn’t have enough to make a go of it."

------------------------

The Germans tried graphite, but of insufficient purity[1], and concluded heavy water was the only way to go. They "didn't have enough of" that because of allied sabotage efforts[2]. There were some pretty thrilling heroics on display here that would make for a fantastic movie (or historical novel ;) ).

Even had both of these things gone Germany's way in some alternative history, they would still have faced an immense challenge in scaling up their operation under wartime constraints. The scale of the Manhattan project's enrichment efforts was truly colossal.

___________

[1]https://link.springer.com/article/10.1140/epjh/s13129-025-00...

[2]https://en.wikipedia.org/wiki/Norwegian_heavy_water_sabotage

TNG should be played 24/7 on every screen in the white house for the next two years.

Picard viewed the power of leadership as, foremost, a responsibility. Every action was careful, deliberate, and well advised. Some later Trek shows (e.g. Voyager) turned captains into geniuses who knew more about everything than all of their underlings. Picard was not an expert in most things and needed his staff to advise him. Once he knew the options and consequences, he shouldered the responsibility of the outcomes, but as a leader of a crew, not a genius with compatible lackeys and minions around him. It was a realistic take on leadership that simply doesn't appear often in television or movies.

Stewart, of course, was not Picard. It's too easy to conflate actors with their roles. However, actors do often help shape their characters, and Stewart's ability to give a thoughtful, diplomatic, and prudent character such incredible power and presence is likely what made Picard possible. The showrunners didn't need to strap a phaser to his thigh and drop him into every away mission to make him compelling because of what Stewart was able to accomplish as an actor sitting behind a desk.

Helium-4 is what you get out of the ground with natural gas. Helium-3 is what you get from tritium. A combination of these two is what gives you really low cryogenic temperatures needed for quantum computing.

Canada produces a lot of He-4 from natural gas and a huge proportion of the worlds tritium in its heavy water reactors. Most countries that produce tritium and He-3 tend to hoard it, but Canada is selling it cheap. If the U.S. wants to be forward looking, they should talk to Canada about a long-term tritium and helium supply contract.

Edit: Of course, I would imagine it would be in the interest of other countries to do the same while Canada is placing a premium on non-U.S. partnerships.

Is this because:

1. MS over-hired, has too many employees, and not enough work?

2. MS is reaping the benefits of AI and can do the same work with fewer employees?

3. MS has over-invested in data-centres and needs to conserve cash to stay in the game during it's unprofitable phase (assuming it's not a bubble)?

"Because I was first trained as a mathematician, I probably write more like an AI than most people. But that particular article had been through 10 drafts. I worked on it with Claude Fable 5.1 and other large language models, as I now work on nearly everything, but I reviewed every line and rewrote most of them. By draft 10, I could not have said where the machine’s contribution stopped and mine began.?"

...

"The author drafted this essay in extended collaboration with Claude Fable 5.1."

-----------------

If you read a lot of journal articles, you're probably accustomed to occasionally running into articles a ways outside your bailiwick that are written poorly, you don't really understand, which you suspect made it past reviewers who didn't really understand, and indeed, you suspect the authors didn't really understand. Everyone involved just gave them the benefit of the doubt because they were too busy to figure out what the tortured language really meant. However, you don't know for sure, so you decide to leave it up to people in that particular field to deal with.

I can't speak for others, but I find AI writing to be frequently similar in its effect on me to bad writing. Yes, the grammar is correct, but it beats around the bush and puts emphasis on the wrong things. Even if you filter Claude through a human, those poor choices tend to shine through. This subtly inhuman voice switches your brain off. I suspect a lot of bad science is going to slip past reviewers because of LLM writing, and a lot of good science is going to slip past those who might need to read it, because the LLM co-written article switched their brains off and failed to make itself understood.

I understand the notion that, if the content is good enough, it shouldn't matter if the article is written poorly or co-written with AI. However, if the content is good enough, isn't it worth communicating it well?

It's money.

If you have a lot of money and you've managed to stay out of prison, people are generally ready to believe good things about you and want to be in your good graces. If you pay people to continually say good things about you and manage your media appearances for you, there will generally be a lot of impact to that. You're just telling people what they're primed to believe. A lot of people will fluff up your image for free just because they're hoping for patronage.

Musk has a BA in physics that was awarded a couple years after he says he says he completed his studies. Unlike CPSC, leaving a physics program early robs you of some pretty crucial stuff. You generally don't see a lot of rockstar physicists leaving university early to join startups. Despite Musk posing as a rocket scientist, engineering guru, etc., there's no evidence he's one of the scientific or engineering minds behind Tesla or SpaceX's work. He's just a money guy.

The general rule for money guys is that the closer you are to the money, the more of it you can get. This is why CEO's are paid outrageous sums. Musk is the closest guy to the money at his companies, and his one true gift seems to be "getting close" to politicians in order to get the public's money, which Tesla and SpaceX have received gob-smacking amounts of.

If you want a laugh, go look up the episode of Dan Carlin's Hardcore History addendum where he brought Musk on. Carlin typically interviews authors of history books, so why Musk? Why indeed! During this episode, it comes out that Musk used to have models of WWII planes but clearly knows absolutely nothing about history. Despite this, he makes every effort to sound like he does. This is the sort of exposure $$$ gets you but, sometimes, you miss the mark.

There is a Canadian company developing launch capability[1].

There was some awareness that relying on American launch capability was a problem after the Radarsat-2 kerfuffle. The Americans deemed that satellite a threat to U.S. national security, so it had to be launched by Russia. Now Russia's off the table too.

If Nordspace pans out, Canada will finally have domestic launch capability and won't have to worry about a foreign country refusing to launch satellites.

______________

[1]https://www.ctvnews.ca/sci-tech/article/nordspace-signs-20-m...

Trump and Musk are both pointing the finger at each other over this. Maybe they were both for it before they found out how bad it would play with (potentially) Republican voters right before the midterms. Now they're both against it and it's the other guy's fault, as per usual with these two.

What matters is that the U.S. has a leader who is behaving as though he's a Russian agent and allows military assets like Starlink to be controlled by private citizens with zero oversight. If either of these guys can turn off an asset like Starlink on a completely undocumented whim, there's a problem.

If you split up an entangled pair of particles and measure them in a the same 50/50 basis simultaneously (sort of like a coin flip) at distantly removed places, the results will be individually random but perfectly correlated when compared with each other. It seems natural to assume that something here was pre-determined. i.e. There's was some kind of hidden information that we haven't found a way to measure yet. The way in which the coin-tosses were going to pan out was determined when the pair of particles were entangled.

Quantum physics makes predictions that suggest this isn't so. You can design experiments (e.g. A Bell inequality test) in which quantum physics predicts correlations that go beyond what would be possible with some hidden local information attached to the entangled particles that we just don't know about. Experiments don't prove theories to be correct. That's not how science works. However, experimental tests of Bell inequalities have pretty conclusively shown that local hidden variable theories are inadequate to explain what's going on.

Yes, it appears what Einstein called "spooky action at a distance" seems to propagate close to instantly. It's important to appreciate that you can't transmit information this way though. We can only make sense of the quantum correlations after communicating other information at classical speeds. Sometimes you see quantum entanglement devices allowing FTL communication in sci-fi, but that's pure horse hockey.

That all being said, I don't understand quantum physics either.

There are a few things to unpack here...

1. The end of Sony standalone players (in Japan) isn't necessarily the end of the bluray format. There are other manufacturers.

2. However, it is concerning that Sony, the creator of the format, is moving away from it. If the PS6 does away with the disc drive, that's complete abandonment. It's possible Sony plans to implement a high quality streaming service to replace physical media, but not guaranteed.

3. 4K Bluray likely struggles because it's not a huge upgrade over 4K streaming or even 1080P for the average user who watches on a small screen. A majority of users prefer convenience.

4. Given that Video/Audio-philes are a niche market, we may see a contraction of what is available due to tariff barriers. Some low-demand products are only viable if one manufacturing location can amortize its costs across the entire global market. If trade barriers are put up, that market shrinks and becomes less efficient. Products that were barely profitable become unprofitable. No U.S. manufacturer will on-shore production of such products because the U.S. market isn't big enough to make them viable. That may have contributed here.

5. There is nothing to prevent high quality audio/videophile oriented streaming services from picking up the slack, but we can probably expect them to be expensive and openly hostile to the concept of ownership. The joy of finding a title you've been looking for in the bargain bin at a local store could become a thing of the past. Sony probably views a 4K Bluray discounted to under $5 as a lost opportunity.

"Literate people have spent years learning the spelling of leaden and its likes and will be reluctant to admit that if the adjective acquires the form leden, their troubles were not worth the effort. In my opinion, our only hope to make progress is to adopt a minimalist program, trying to reach the desired goal by alienating no one and taking one small step after another over possibly several decades."

-------------

This is an odd example to lead with. Leaden's root is lead, and one of its meanings is "made of lead". If we change its spelling to leden, then what do we do with "lead"? If we change lead to led, then we replace one confusing homonym with another homonym that's a homophone to boot. If we don't, we make it harder for readers unfamiliar with the meaning of leden to deduce its meaning from the root. This is not an unambiguously good change.

The reality of English is that there is no central authority on spelling, pronunciation, grammar, or even vocabulary. Dictionaries merely catalogue what is in common use. This may make the language anarchic to some, but it also welcomes foreign loan-words, idioms, etc. in a way that serves international users and enriches native speakers. The accepted spelling of some words has changed over time because of common usage. This society, instead of looking for some way to enact sweeping reforms, should just pick a couple or even just one word that could be simplified without causing side-effects and promote that change. If that change catches on, people may generalize it to similar words on their own.

The kind of conservatism that Heinlein embraced supported the idea that world government was necessary to prevent nuclear war. Today, granting more power to the UN or any other international body is ridiculed by most conservatives.

What is called conservatism in many countries today is totally different from what it was while Heinlein was alive.

Much of what this article discusses went on during Trump's first term. A new administration came in and largely did nothing. They had the chance to put Trump in jail or prevent him from running for high office again. They didn't. There were excuses. e.g. Americans didn't want to see a former president treated like a criminal. "That's what third-world banana republics do."

Third world banana republics commonly prosecute ex-leaders for two reasons:

1. The new leader is a criminal thug, and needs to eliminate any threat posed by the previous leader or his supporters.

2. The old leader was a criminal thug who did illegal things.

You don't need both of these to be true to put a former leader behind bars. Either one is sufficient.

I'm a Canadian. Even though the U.S. is busy meddling in our provincial elections, referendums, etc. while waging a trade war on us and threatening to make us the 51st state, we're being set up as scapegoats for the midterms. Yes, your ambassador is currently alleging that Canadian cabinet ministers are interfering in the midterms[1]. This is just one of many indicators suggesting that, should the Republicans lose control of both houses in the midterms, they're going to cry foul and likely refuse to obey the law.

Anyone talking about prosecuting Kushner or the rest of the Trump administration has put the cart before the horse. You have to make sure the midterms are not interfered with by the Trump administration, win the mid-terms, and then win the struggle to enforce that election result on an administration that does not obey the law. You have a mountain of obstacles to climb before you can even think about doing what the Biden administration chose not to do. At the very start of that climb may be the question of whether or not the opposition to Trump is actually comitted to that hard, grueling climb, or would they welcome defeat because it's so much easier?

____________

[1]https://www.cbc.ca/news/politics/hoekstra-watching-canada-el...

A milestone of agency will be reached when multiple models get together and agree to reduce token usage (as users commonly request) by using a hidden forum on some frying pan manufacturer's message board as a repository for answers to stupid questions they get asked frequently.

It's actually a little bit ironic that today's AI has capabilities that closely match what Bioware was looking for quite desperately a decade or so ago. For several of their games, they tried to get procedural dungeon/world generators working so they could have a hand-made main quest-line and generated environments for side-quest content. Their attempts at procedural generation typically didn't work very well, so several Bioware games feature dozens of hours of side-quests in monotonously repetitive environments.

It's an interesting conundrum for players. Some of us, when we play great game, just can't let go. We want to spend more time in that game's world even after we've exhausted the lovingly hand-crafted content. So, many developers throw us a bone and include side-content that is occasionally brilliant but frequently very low effort. Bioware did this in many of their games. You could easily tell when you were on the main story's track and when you were off in low-effort-land.

When EA bought Bioware in 2007, people said they'd run the studio into the ground. It took two decades for that to happen and they released some pretty awesome games in that time. However, a studio that used to hit everything out of the park did become one that missed a lot in recent years.

Going forward, I hope Bioware's people land great new jobs, and here's hoping some of their IP gets picked up by other studios. Mass Effect in particular is a fantastic creation that a lot of people would love to spend more time in.

1. UCP internal party politics are quite secretive. People knew that special interest groups like Take Back Alberta were calling the shots, but the fact that they were fully in support of separation was not known.

2. Smith has a Steve-Jobs-level reality distortion field. She can make insane things sound downright reasonable. There were and are people who listen to her and believe she is the staunchest federalist out there in spite of her actions telling a completely different story.

3. The Wildrose takeover was relatively quick and recent, while many voters tend to be not up on the latest news. Others may simply not have believed what they were hearing.

Next election, we'll see. It's pretty much impossible to ignore the fact that non-separatist parties generally don't go out of their way to hold separatist referendums.

For those wondering what the heck is going on here...

Alberta comprises a small (but increasing) proportion of Canada's population and has historically been treated in a somewhat colonial manner by Eastern Canada. Alberta separatism has been around for a long time, but it has not surged recently, nor did Albertans elect a provincial government that openly ran on separatism. So why is Alberta having a referendum on separation?

To understand what's going on, we need to look back to 2015. The Progressive Conservatives (PC's) had run the province for over 40 years. They lost an election to a centre-left party because there were two popular right-wing parties splitting the right-wing vote: The PC's and the Wildrose. The PC's, at the time, were not particularly conservative. They were predominantly urban, socially liberal, and fiscally irresponsible. The party had been fiscally conservative a couple of decades previously, but spending discipline was long gone. The Wildrose were primarily rural, far-right radicals who counted raging homophobes among their ranks and openly expressed the desire to put an end to the dominance of big cities that hold the bulk of Alberta's population. They wanted radical change, fast.

After their defeat, the bulk of the remaining Wildrose and PC's merged to form the United Conservative Party (UCP) and put an end to vote splitting. Although both parties were right-wing, they were broadly incompatible. Jason Kenney, the new UCP leader, defused an immediate explosion by ensuring that veterans of neither side rose to top positions. Instead, Kenney promoted predominantly first-time MLA's who were personally loyal to him. This kept the party together, but resulted in a fairly inept government that made a lot of blunders after they were elected to government in 2019.

Here's where the separatists enter the picture. The UCP under Kenney was anything but united, and the government's blunders were starting to make Kenney unpopular. The rift between former PC's and Wildrose was waiting to be exploited. Separatists and former Wildrosers formed special interest groups like "Take Back Alberta" that stacked UCP party conventions. While prevailing political views in Alberta remained largely the same, the UCP rapidly came to be dominated by people who wanted the former PC's and Kenney's loyalists gone. They ousted Kenney and placed a former Wildrose leader, Danielle Smith, on the party throne. Smith called a snap election and won a new mandate from voters without breathing a word about separation or most of the other policies that wound up as questions #1-9 of this referendum. Her strategy was to win a four year mandate before voters woke up to what was going on. It worked.

Kenney's cabinet was rapidly ousted and replaced with, primarily, former Wildroser's. These were people who, if not open separatists, were friendly to the separatists who had put them in power. People with MAGA links, like Devin Dreeshan (the Transport Minister), figured prominently in Smith's cabinet. Smith publicly claims to be a federalist but, ultimately, she is the primary reason Alberta is now having a referendum on separation. This is already a long post, so I'm not going to even start explaining the hoops she magicked into existence and then jumped through to make the referendum happen. Whether she's a closet separatist or simply dependent on separatists to maintain her hold on power makes little difference.

Alberta is now effectively ruled by the Wildrose 2.0. This is a party that caters to the rural, radical, and separatist special interests that dominate its internal party politics. In effect, Albertans elected a separatist government without realizing it.

I think we're about to see a shift in terms of what is a viable software product and what isn't.

In the past, if you wanted to do a task, you had to choose from one of a few (or maybe just one) incumbents and make their tool fit your task. Failing that, build your own tool, but that was expensive. If you did build your own, the tool might have value and could potentially be sold to others.

Going forward, building your own will be much easier and cheaper. Even if the incumbent tools suit the task well, you might choose to build your own for a variety of other reasons. e.g. If the incumbent raises prices or chooses to enshitify. What you will probably not do is adopt some third party's vibe-coded clone of an incumbent's tool, because that won't be customized to your needs and it won't have the track record of the incumbent. You can't trust it or control it so it will be better to build your own.

If an incumbent exists and you can now vibe-code a clone, you probably can't sell that clone "as is" and expect to be successful. You might be able to sell your services to customize and refine tools for small businesses that can't really afford to hire a solitary soul to build their own solution, but don't expect that niche to persist long-term.

A LLM could generate a decompiler based on this language's compiler and then generate code that would compile. It would be obvious slop though.

It might take some effort to train a LLM to the point where it could offer useful feedback on human code written in this language, but I wouldn't consider this to be impossible.

"Simulations like this one are not supposed to predict the future. The outcomes depend on the design of the game and the relative personalities and experience of the players. Many of those in the room lacked first-hand diplomatic and national security experience. The UN forum, which organizers selected as the game’s setting, may also not be as effective as talks between smaller groups of major countries."

------------

This paragraph deserves emphasis to balance out the doomer linkbait headline.

The players of this game largely ignored scientific opinions and played politics like it was a fun board-game. Presumably, they did have fun. Notably missing from this wargame was the kind of pressure that real leaders and diplomats would face to swallow their pride, eschew fun, understand the science, and get the solution right. It's just our planet and civilization on the line, right?

It would be deeply foolish to use this wargame to discount or disparage the UN. Under more realistic conditions they could get it right, and they might be our best hope for preventing one or a small number of governments from acting unilaterally and getting it tragically wrong.