I think the issue people run into is that the types of errors the LLM makes are ones humans wouldn’t make, like having a basic misunderstanding of the goal of the software and making silly logic errors that technically work, but don’t serve the correct purpose.
This can be seen more clearly with self-driving cars as an example. A self-driving car may be safer than a human driver, but when the self-driving car plows into the side of a semi truck in broad daylight… that’s generally not a mistake a human would make. Humans and AI have different failure modes, so when AI fails where we generally wouldn’t, it really stands out and gets judged harshly.
The iPhone has a mode called Assistive Access. It’s meant for older people or those with cognitive impairments, but could work well for children as well. It changes the UI, limits the apps available, and simplifies those apps, so they just have very basic features.
You could set it up on a cheaper or hand-me-down iPhone (if you have one).
Very few people buying a new car keep it long enough to make it cheaper than buying $5-10k cars.
The last cheaper used car I bought, I kept for about 10 years, it was $10k. I sold it to my brother in-law after that, who had it another 5 years… it would have gone longer, but he didn’t realized I had changed the timing belt before I sold it to him, so he misjudged how long the one he put on would last, and it destroyed the engine… anyway, let’s go with it..
The last new car I got, was relatively inexpensive, as far as new cars go, at $35k. If a $10k used car ran for 15 years, I’d have to keep that $35k car for over 50 years to make it worth it.
New cars are almost never financially justifiable. And that new car will run into the same maintenance issues and costs as the used one as it ages.
The nicer the car, the worse this equation gets, as increased car pricing translates to luxury, not quality. This makes them very different from work boots.
The question isn't so much how long it lasts until the engine explodes, it's how expensive is maintenance. That's the boot theory part. If you can't save up enough for the $10k car that lasts 15 years (though you mentioned a timing belt replacement and on some cars that's a few thousand $, ask me how I know...) and instead can only afford to keep buying $1000 cars, that could turn into a slew of smaller fixes that add up to a lot more than the car was initially bought for.
I'm also a bit surprised you'd buy a new car. Off-lease is much cheaper and they're still fairly new ish.
I hear this a lot, but it’s not an excuse for never starting. Amsterdam started taking their non-car infrastructure seriously in the 1970s and now it’s seen as an example for the rest of the world.
“A society grows great when old men plant trees in whose shade they know they shall never sit.”
Asking to see a realistic roadmap and timeline is being a realist, not a defeatist. If Amsterdam started in the 1970s, that’s fifty years ago. Maybe we can make progress by the 2060s-2070s? Or we can just wait for China to manufacture all the EVs the world needs. Which is more likely to occur first?
Consider the US has $40T in debt and no will to raise taxes at this time, and therefore no will to spend on public infrastructure projects. Where will the capital come from?
How and when will we start? How many elections need to occur to appropriate funding? How much of the US electorate has to turn over for the right people to be elected to appropriate funding? Until shovels are in the ground, it’s vaporware and dreams.
It doesn’t have to be a massive nationwide project on par with the national highway project. It can start with local communities.
The mayor of Carmel, IN thought traffic circles were better than intersections and wanted more mixed use development. They have that now, nearly every intersection in the city has been replaced over time as they were due for routine work.
In Detroit they are in the process of building out a 29 mile greenway around the city.
These city level projects are a lot more important than the national ones, when it comes to reducing car traffic in cities and having an impact on people’s most frequent trips.
Seeing this at a local level could inspire more rail connections between cities, which could also come from the private sector. We have already seen a couple examples of this.
For anything funded by the federal government, you’re right. They have a huge mess on their hands, and need to get serious about cleaning it up. They have no business kicking of any major public works projects until that happens (though that has never stopped them in the past). But at a local level, things can still move forward and we are seeing it happen to some extent, though it often still feels more like a hobby than a firm conviction.
I know they said the A20 was desktop class, but beating the M5 is wild to me. I can’t wrap my head around the fastest chip (at least for single-core performance) is in a phone instead of a desktop.
I thought maybe they clock the M5 down for the iPad, but when I go over to the Mac benchmarks, I see the same thing.
You don’t need to upgrade every year, they are simply available. The yearly upgrade rarely makes sense, as things don’t change much year to year. However, when you are ready for something new, it’s nice that a recent phone it’s available and you aren’t having to buy something that came out 4 years ago. Also, over time, those yearly updates stack to turn into a pretty meaningful improvement over time.
> You don’t need to upgrade every year, they are simply available.
Not only that, in the last half-decade or so I've actually started viewing updates as a chore, not an upgrade.
I couldn't imagine wanting to upgrade phones annually anymore. You have to order the phone, receive it, transfer and set it up, there's potentially SIM card changes (although I'm now on eSIM), and then dispose/sell of your old phone (where did I put the box, how much am I going to sell it for, etc).
This is why they went on to say it doesn’t make sense with modern RAM sizes.
I remember when I first came across this rule when getting into Linux for the first time. If memory serves, I had 256MB of RAM, so I set swap for 512MB, which seems perfectly cromulent.
I regret signing up for 98% of the sites I’ve signed up for over the years, if for no other reason than deleting those accounts (and the data attached) is difficult or impossible.
This can be seen more clearly with self-driving cars as an example. A self-driving car may be safer than a human driver, but when the self-driving car plows into the side of a semi truck in broad daylight… that’s generally not a mistake a human would make. Humans and AI have different failure modes, so when AI fails where we generally wouldn’t, it really stands out and gets judged harshly.
reply