skohan

I wonder to what extent the 4o rollback was motivated by this exact case

I had the same experience. For me peanut butter and apples were my "healthy snack" and were accounting for a huge amount of calories per week.

Not only forests but even just tree cover. I live in Berlin and there were a few days a few years ago over 40 degrees, and I remember the difference was stark going from my relatively tree-lined street to an area of the city with only concrete.

I always felt unwrap is preferable for diagnosing issues, which makes it useful in tests. A failed unwrap will point you directly to the line of code which panics, which is much simpler than trying to trace an issue through many layers of Results.

If you use `assert` in tests, I don't understand why you wouldn't also prefer unwrap in this context.

I think it's also perfectly reasonable to use in a context like binary you run locally for private consumption, or for instance a serverless function which allows you to pinpoint the source of errors more easily in the logs.

It's not a good fit for cases where you do expect cases where the unwrap will fail, but it's a great fit for cases where you believe the unwrap should always succeed, as it allows you to pinpoint the cases where it fails.

I was a huge Swift fan but SwiftUI and the changes supporting it in the language got me to switch to Rust for all my personal projects

Doesn’t rust’s static linking also have to do with the strategy of aggressive minimization? Iirc for instance every concrete instantiation of a dynamic type will have its own compiled binary, so it would basically be impossible for a dynamic library to do this since it wouldn’t know how it would be used, at least not without some major limitations or performance tradeoffs

I believe MS is trying to standardize this, in the same way as they do with DirectX support levels, but I agree it's probably going to be inherently a bit less consistent than Apple offerings

There was a rumor floating around that Apple might try to enter the server chip business with an AI chip, which is an interesting concept. Apple's never really succeeded in the B2B business, but they have proven a lot of competency in the silicon space.

Even their high-end prosumer hardware could be interesting as an AI workstation given the VRAM available if the software support were better.

I don't understand why you would need custom firmware. It seems like you could go a long way just implementing back-ends for popular ML libraries in openCL / compute shaders

I love the air form factor. I do serious work on it as well. I have used a pro, but the air does everything I need without breaking a sweat, and it's super convenient to throw in a bag and carry around the house.

a dedicated Nvidia rig

I am honestly shocked Nvidia has been allowed to maintain their moat with cuda. It seems like AMD would have a ton to gain just spending a couple million a year to implement all the relevant ML libraries with a non-cuda back-end.

I use Zed as a daily driver, and I haven't once used the collaborative features

In fact they probably care that it's not easily portable

Molten VK is amazing. When I started working with it, I was expecting a lot of caveats and compromises, but it's shockingly similar to just using Vulkan that you can easily forget that there's a compatibility tool in play.

Probably you can squeeze a bit of optimization out of using Metal directly, but I think it's a more than viable approach to start with Vulkan/MoltenVK as a target, and add a Metal branch to the renderer when capacity allows (although you might never feel the need)

It’s become “grind leetcode” which is clearly a failure mode.

I have no insider knowledge, but I always wondered if those kinds of cultures had a hazing / blind-leading-the-blind aspect to them. I.e. the people who got hired were the ones who jumped through some arbitrary hoops, so they doubled down on those hoops being the right hoops to choose the best candidates.

Unfortunately I think "they share all the same apps" will not include a terminal with root access, which is what would really be needed to make iPad a general purpose computer for development

It's a shame, because it's definitely powerful enough, and the idea of traveling with just an iPad seems super interesting, but I imagine they will not extend those features to any devices besides macs

This is no doubt a milestone we should celebrate. At the same time, the Brazilian beef industry is the driving force behind slash-and-burn expansion of agricultural land into the rainforest, so Brazil is still a massive contributor to climate change.

I guess the reason it strikes me as strange is, on linux the usage of `sudo` helps you understand which operations might be dangerous in a fine-grained way

So like if you are executing a series of commands, the one requiring admin privileges tends to be one you might want to be more careful about (i.e. altering system configs, or doing a potentially insecure operation)

So if you are running everything in an admin terminal, it seems like you wouldn't have that extra check to remind you to be extra mindful of a particular operation, since everything you do in that terminal is in the same bucket

Yes terrible move which will lead to much confusion and consternation in the future.

But honestly I'm most amazed by the fact that there wasn't previously a way to run commands with elevated permissions in Windows. How did people work like that? Just run everything in an admin terminal super unsafely?

I'm sorry Dave, I'm afraid I can't do that

Imo it should give you the answer, and make you aware of the risks.

Some people won't follow the advice, but that's what code review and static analysis tools are for.

God I hope we don't end up in a world where LLM's are required to take an ideological stance on things like software architecture or testing paradigms

I don't live in the US anymore, so it's hard for me to speak to the US healthcare system, but as an outsider it seems to me that the problems with the system stem from the incentive structure, and won't be fixed by any technology.

It seems to me the US healthcare system sits in this weird gap where it's not funded by private individuals, but it's also not publicly funded (excluding medicare/medicade/VA/tricare etc)

So it doesn't benefit from efficiencies of a single-payer system like the UK has, and it doesn't have to obey proper market forces either, since you have a bunch of for-profit entities in the middle, but the costs are obfuscated from the consumer unless you're unlucky enough not to have employer-funded health insurance.

So until you fix those core issues, I don't see how technological advancements are going to have much of an effect either positive or negative.

It's an orthogonal issue.

Yes I understand that, but the problem is there's no structure to enforce decisions at the level of human civilization. So even if something is agreed and enforced by 99% of jurisdictions, that doesn't mean it won't happen

The problem is who is "we"?

If the capability exists, someone will pursue it. Maybe the G20 will pass a resolution to put safety measures in place, but whoever decides to flaunt those safety measures stands to gain an asymmetrical power advantage

That sets up a competitive dynamic where players are incentivized to try to get there first, no matter the risks, because if they don't someone else will

Similar dynamics played out with the nuclear arms race

But you don't need AGI to be dangerous right?

all signs seem to be pointing toward a longer timeframe rather than a short one.

Not saying we are closer, but isn't the risk being described that we could pass an inflection point without knowing it, and things could start accelerating beyond our control?

Human beings tend to think about things linearly, and have trouble conceptualizing exponential growth.

Yes both of which could be helped by augmenting the healthcare system with AI to relieve practitioners from repetitive, menial tasks

due to lack of training sometimes they miss things that an MD would have caught.

It goes the other way as well. My aunt is an RN and she loves to tell the story about the time that a doctor was in such a rush one time she had to call his attention to the fact that his patient was dead.

So for example, a treatment plan might look like this:

1. Go to a GP and explain symptoms 2. GP orders some tests based on symptoms 3. GP refers you to a specialist based on test results 4. Specialist orders more tests 5. Specialist analyzes results and orders more tests, or prescribes treatment plan 6. Check in with specialist in weeks or months to evaluate results and update treatment

It seems to me that for many common issues, 1-4 could plausibly be replaced by AI guided care with minor downsides.

If you are normally spending a lot of time waiting between steps for a practitioner to become available, then it's going to be a net win for many people.

Presumably you would structure such a medical system such that lower level practitioners would know enough to know when to escalate to an MD.

90% of medicine is stuff like people going to the doctor with the flu to get a sick note to stay home from work and instructions to rest and drink fluids. I don't think you need an MD for that.