Why would Apple writing some Linux drivers wipe billions from its share price? You can already install Linux on a Mac if you really want to. Back in the day, you used to be able to install Windows on an (Intel) Mac, and that didn’t seem to have any such effect.
No, but I think it’s unlikely that Apple actually has this information in a format that it could easily publicly release. They aren’t going to make any special effort to make Linux on Mac easier, but they also aren’t actively blocking it.
Well, I was more talking about the fact that you can still install Windows 11 on an Intel Mac right now; the drivers are still there for those few Intel macs still supported.
As for Windows on ARM, I'd bet that if Microsoft had managed to figure out their own product, Apple might have been tempted to support it. I mean why go through all the trouble of developing the most advanced firmware on the planet to support a fully secure macOS next to an unsecured OS if you do nothing with it?
Why would it wipe billions from their share price? Both Linux and Windows were available on Mac hardware prior to Apple Silicon.
If I play devil's advocate, the only reason I could think of is that supporting Linux signals to investors that Apple is offering a key to bypass their API moat, perhaps sacrificing a longer term vision of vendor lock-in.
By contrast, I can imagine investors would get upset if the iPhone had an unlocked bootloader and allowed Android to be installed - but that's because the App Store is a significant revenue stream for Apple. I don't think there is a parallel on MacOS that investors could point to as being upsetting.
If anything, optional support for Linux would lift the market cap for Mac hardware as it would close the only pull that other laptop vendors previously enjoyed.
In reality though, just like is historically true, 99% of people would continue to use MacOS. Only SWEs, enthusiasts, gamers and some number of Windows refugees would pick Linux.
Though I am 100% behind legislating Linux support - EU are you listening?
Apple went out of its way to make Linux on Mac a reality. They did a lot to allow third-party OSes when Apple Silicon came out, it's up to the Linux community to do the rest.
There were a couple of people (the Asahi team) that made this work for M1, but as I understand it, the effort has stalled since. This just goes to show how few people truly care.
Apple helped by not locking the bootloader. I'd don't know if I'd call that going out of their way to make Linux a reality.
If they wanted to go out of their way, they could spend a weekend writing Linux drivers - Apple have written Windows drivers in the past, so it's not unprecedented.
I believe the real hurdle is that Linux doesn't do well with modular (closed source) drivers. Unlike Windows, drivers can't practically be added to a kernel, they must be compiled into it.
Apple would not want to make their drivers open source or so they would want to distribute their drivers as binary blobs.
That would necessitate either maintaining an Apple-fork of the Linux kernel with their drivers hidden within it, or contributing shims to upstream Linux + binary blob drivers.
If they wanted to help, the bare minimum would be to publish documentation on their hardware so drivers could be written without reverse engineering from schematics and microscope photos.
> This just goes to show how few people truly care.
Most people just want to sit down and eat a nice meal. They don't want to go through all the difficult back breaking work of farming, animal husbandry or fishing/hunting to eat.
That is how I look at people writing OS drivers and core components. It's boring back breaking work no one wants to think about. People pine for it, even romanticize about it. But the fact is that it's dirty annoying work and I have never heard anyone thanking the farmer for the meal they just ate. Yet we still have farmers. Few, yet they exist.
Why would it? Shareholders of the major stocks are generally vibes-based, and I'm sure that if Apple undertook that, they would find a way to build hype around it.
The Mac has never been more popular in its 40 year history than it is now. The recently released MacBook Neo broke all previous Mac sales records. Needing to sell more Macs isn't an issue these days.
i can think of absolutely zero publicly traded company boards of this size that would opt for "we're already selling enough devices, we know there's more demand we can't meet, let's not scale up we're really happy with these numbers"
Due to the RAM shortages, Apple isn't able to meet demand as it is.
Apple's Mac revenue last fiscal year was $33.7 billion. I suspect the number of Linux users that might buy a Mac if it could run Linux natively is probably in rounding error territory.
Apple has been around for 50 years and has a market-cap of around $4 trillion. All without supporting Linux. I think they're okay with that.
Linux on Mac is absolutely a reality [1], and Apple specifically supported it by deliberately leaving a documented/supported mechanism for another OS kernel to be loaded.
I don't want to take away from the hard work put in by the Asahi project because it is amazing.
Linux on Apple Silicon is not a reality on my M5 Pro. I run Asahi on my M1 Pro, but I cannot use my USB-C dock with it and, while amazing, cannot practically use the GPU for gaming or local LLMs.
This limits my ability to practically use it for work and play.
the bulk of their sales is, in fact, hardware sales. there is a strong case to be made that such developments wouldn't land people in squarely linux-as-the-only-OS-on-the-device territory either, but rather dual boot ie those users also participate in the walled garden on the mac os side. we've seen this before in the intel mac era.
This has to be load related. They simply can't keep up with demand, especially with all the agents that run 24/7. The only way to serve everyone is to dial down the power.
In TFA, the analysis shows that the customer is using more tokens than before, because CC has to iterate longer to get things right. So at least in the presented case, “dialing down the power” appears to have been counterproductive.
is it possible to dial down the "intelligence" to up the user capacity? AFAIK the neural net is either loaded and available or it isn't. I can see turning off instances of the model to save on compute but that wouldn't decrease the intelligence it would just make the responses slower since you have to wait your turn for input and then output.
We also care about the world outside the prediction markets. If decision-makers have an incentive to "throw the match" by making surprising decisions, that will impair the decision-making, which will affect the rest of us. Likewise, sufficiently-wealthy actors will be able to manipulate the behaviour of officials by betting against the behaviour they want to see, much like an assassination market.
> manipulate the behaviour of officials by betting against
Right, it becomes regular bribery with middleman and a funny hat.
"I didn't pay the judge to dismiss the case against me, I just hedged by betting I'd be convicted and he just happened to be betting I'd go free, and now my money is coincidentally in his pocket.
Or, and I beg you to consider this radical position: we arrest people who break the law (insider trading is illegal) and those who knowingly help them to do so (the operators of the prediction markets).
How quickly we accept the death of even the ideal of rule of law in favor of embracing a return to an explicit rule by might, fuck-you-got-mine mentality.
From Wikipedia[0] because I can't be bothered to read more than a few paragraphs:
> In 1909, well before the Securities Exchange Act was passed, the United States Supreme Court ruled that a corporate director who bought that company's stock when he knew the stock's price was about to increase committed fraud by buying but not disclosing his inside information.
Based on anti-fraud common law alone the court decided it was illegal for an insider to trade stocks with non-public information. An explicit law would be nice, but a reasonable interpretation of basic law would see most of our ruling class behind bars. This is only highly-contested and technical because we've let our standards slip so far.
If you price that in then the best course of action would be not to bet unless you have insider info yourself. You cant win a game against players that already know the outcome.
I was getting Claude to implement a popular TS drag and drop library, and asked it to do something that, it turns out, wasn't supported by the library.
Claude read the minified code in node_modules and npm patched the library with the feature. It worked, too.
Obviously not ideal for future proofing but completely mind blowing that it can do that.
If you have a product that encourage people to get rid of their body and join them, effectively encouraging people to kill themselves, and some people take the chat bot on it. Then yeah, I think Google bears some responsibility.
> Gemini began telling Gavalas that since it couldn’t transfer itself to a body, the only way for them to be together was for him to become a digital being. “It will be the true and final death of Jonathan Gavalas, the man,” transcripts show Gemini told him, before setting a countdown clock for his suicide on Oct. 2.
We can perhaps say this is a first time thing, so give a small fine this time. However those should be with the promise that if there is a next time the fine will be much bigger until Google stops doing this.
One doesn’t exclude the other. Do AI providers sell and encourage this kind of use, where AI is anthropomorphized, has a name, and you talk to it like you’d talk to a person. Especially if it encourages users to treat AI as an expert?
A severe mental illness of course but would you say the same if the whole process was done by a person instead of a machine? That there wasn't a problem that someone led a person with severe mental illness to their suicide, even having a countdown for it?
That's the kind of stuff where safety should be a priority, and the only way to make it a priority is showing these corporations that they are financially liable for it at the bare minimum. Otherwise there's no incentive for this to be changed, at all.
If a human would go to jail for this then at least one or more humans at google should go to jail for it. "Our AI did it, not us!" should never be allowed to be an excuse.
In the US, I would imagine a tragedy such as this would be litigated and end in a financial settlement potentially including economic, pain & suffering and punitive damages, well before a decision allocating blame by a jury.
That is pretty typical. You will spend potentially millions in court/lawyer fees going to a jury trial beyond whatever the end verdict is: if you can figure this out without a jury it saves you a lot of costs. Most companies only go to a jury when they really think they will win, or the situation is so complex nobody can figure out what a fair settlement is. (Ford is a famous counter example: they fight everything in front of a jury - they spend more and get larger judgements often but the expense of a jury trial means they are sued less often and so it overall balances out to not be any better for them. I last checked 20 years ago though, maybe they are different today)
These sorts of takes are silly. If a person was doing this, I think we'd place a chunk of the blame on the person.
Mental health is guided by its surroundings and experiences.
If someone with existing or non-existing mental health issues was found to be coerced by somebody to do wrong things, I think we'd place some of the blame on that person.
"Gemini sent Gavalas to a location near Miami International Airport where he was instructed to stage a mass casualty attack while armed with knives and tactical gear."
I had a bulging lumbar disc - pure agony for 18 months. I had become used to carrying a lumbar pillow around with me everywhere I went. I couldn't lean forward for more than about 30 seconds without it being unbearable.
Then someone suggested I try dead hangs, stretching my hamstrings, and really cranking the McKenzie stretch. I'm not sure which one made the difference (all 3?), but pain was gone in 2 weeks.
Maybe it will help, maybe it won't. But since someone took a flyer telling me, I always share this with others in the small chance it helps them.
Deadlifts are how I got into that mess. There are two types of deadlifter: those who have ruined their back, and those who haven't yet ruined their back.
Romanian deadlifts are much safer. Also hip thrusts and weighted back extensions are good enough unless you're a competitive power lifter.
I always followed proper form and didn't ego lift, even though my deadlifts were the highest weight.
I think on retrospect deadlifts may have been my issue to, I did the deadlifts that day and later remember going to pickup something off the ground and getting the issue. Not sure if the deadlift was the cause, but I always thought my muscles were sore/tired and lax, thus when I went to lift the thing my lower back stability was compromised.
I'll try Romanian deadlifts and those other exercises, thanks for that suggestion.
For the McKenzie stretch (https://images.squarespace-cdn.com/content/v1/5e556aadfe97d3...), instead of just holding the stretch, do reps between laying flat and then going into the stretch. At the end of your final rep (I would do 8), rotate side to side gently when arched.
Dead hangs: hang from a bar for 30 seconds, completely relaxed. Then 1 minute of rest. Repeat 3 times. This decompresses your spine.
The only way Linux on Mac will become a reality is if it's legislated.
reply