realharo @ realharo @lemm.ee

Posts

0

Comments

304

Joined

2 yr. ago

There are many different types of capacitors, most of which don't contain any liquid at all (including the most common type - ceramic capacitors).

But in general, you would use specially rated components and materials if you need them to last decades - not the cheapest most basic parts you can find.

Other types of implanted devices have dealt with the same things decades ago.

Some already give you a 30% "discount" (i.e. the regular price is just higher) if you sign up for an entire year.

It does - it's also mentioned in the article.

At least Android also proactively asks them whether to disable notifications for an app if they always swipe them away, or if they haven't used the app in a long time.

I gave GPT-4 a simple real-world question about how much alcohol volume there is in a certain weight (I think 16 grams) of a 40% ABV drink (the rest being water) and it gave complete nonsense answers on some attempts, and straight up refused to answer on others.

So I guess it still comes down to how often things appear in the training data.

(the real answer is roughly 6.99ml, weighing about 5.52grams)

After some follow-up prodding, it realized it's wrong and eventually provided a different answer (6.74ml), which was also wrong. With more follow-ups or additional prompting tricks, it might eventually get there, but someone would have to first tell it that it's wrong.

You have to go where the people are.

The default now is that apps have to first request notification permissions, on both iOS and Android.

The battery life of a typical switch can easily be like 5 years though.

As someone has to mention and program those robots.

Why couldn't an AI do that?

Someone has to create programs and games, someone has to maintain the infrastructure.

Same question.

Youtube videos and streaming became a job.

This will only work because of the parasocial aspect, and there will probably be strong competition from AI there too.

For every thing you imagine, simply ask yourself - will AI be able to do it better?

So far I haven't heard anything convincing where the answer would be "no".

This whole "giving inputs" argument is 100% leaning on today's technological limitations.

With enough advancements, no input you could ever come up with will be able to compete with the automated ones - even if they are working from some very high level goal, like "make something people want" (to give a slightly exaggerated example).

Nobody's going to pay you to utter the phrase "make something people want" (and it's not competitive as a business either).

You are assuming that progress in AI capabilities will stall somewhere close to its present day state. Because today, a professional-made poster will still be better than one you can make yourself. But that won't be the case forever.

This is more akin to how there used to be elevator operators vs. people just pressing a button themselves, or how people couldn't easily book their own airline tickets without going through a travel agent, and now they just order them through a website.

That will not be a marketable skill, if the intended "customer", who just wants the end product, can do all of that themselves.

There are already improvements being made in understanding the intent better, which will eventually render all "prompt engineering" unnecessary and obsolete.

The necessity to tweak prompts will be a very short-lived thing from these early days. At best it will give you an extra year or so.

Similarly if you picture yourself as an owner of a company - you cannot sell something to people that they can just make themselves with zero effort required. Especially in an environment with a million competitors. At best your moat could be the network effects of a large user base, but that's not an easy place to get to.

I try to learn how to use AI

And exactly which part of this process could not be done by AI too?

Which part will still require hiring a human?

If that's the case, they need to stop with the deceptive marketing. Because they are absolutely outwardly promising career opportunities.

They possibly are (or at least have people doing research), it's just not very good (yet?) https://aimodels.substack.com/p/apple-is-working-on-multimodal-ai

Remember the early days of Apple Maps?

Until what? 100% replacement of human-driven cars? Being rolled out for areas covering 50% of the population? Where is the goal line here?

We are already at the stage of commercial operation, with rides available to the general public - even though only in a few locations.

Sure, it's far from being everywhere, but why pretend that progress has stalled, when it clearly hasn't?

Yes actually (except more than a few years).

Waymo is already operating a robotaxi service in 3 cities, now they just need to expand and find a way to make it not lose money.

The right way to implement this is where they don't even have any persistent identifier that could be used for tracking. They should only ever see a derived single-use signature that after verification gives them a yes/no answer and nothing more.

But what is the actual real-world practical solution for those people?

Are they now just going to be broke (or even in debt) with no place to live?

What about all those unfinished buildings that people got in debt to buy?

Do you have the full text of the notification that you could post here? Kinda hard discussing the specifics otherwise.

If it really contains the quote "Congress is planning a total ban of TikTok", I do consider that misleading.

People here are often making a lot of noise about disinformation campaigns on sites like Facebook and Twitter and YouTube (and that's just from user-posted content that the sites fail to moderate, not posted by the sites themselves), so I don't see why this would get a pass.