CEL is much more computationally limited as it aims to keep evaluations in the microsecond range.

With OPA you can easily create policies that take tens, hundreds or even thousands of millisecond.

That comes at the expense of a lot of power though, so much of the complex logic that you can write in OPA simply isn't achievable in CEL.

[redacted]

That's quite interesting. When do they start to understand that saving is better, if they do? I can imagine kids never wanting to save for later, but also I remember that I enjoyed saving more than spending coins when I was a kid.

Brilliant idea!

Do you deduct short term and long term capital gains taxes?

Most tech startups create a monopoly by hiding real-time access to user data behind closed down apps which can force-include ads in your feeds.

A simple EU regulation that requires offering real-time distribution of user data to third parties would massively level the playing field.

Or you can just add some randomness to the prompt by adding “Your random seed is mciifjrbdifnf.”

I just tested that and got 4 left and 2 right so it works pretty well.

As they say... every exponential is a sigmoid in reality.

Human assistants might also just be worse than AI assistants.

When I received my first (and only) BeMyEyes request I spent the first minute or so figuring out how to work the app and the video delay.

“Let me turn the volume up” “Oh wait it’s on my AirPods” “Could you move it more to the left?” “No your left.” “No; not that far”

I’m quite confident that I was a less than optimal assistant and an AI might’ve well done better than me.

It sounds like as part of the sign-up they should ask an experienced helper to make an "artificial" request of the new signee, to give them a chance to see the workflow and to provide an opportunity ask questions of someone who's been in the role that they've signed up for. If there's really an excess of helpers, generating one extra request per new future helper seems reasonable.

I’d guess that they have a rating system and prefer assigning request to helpers who reliably perform well… which might be why I didn’t get repeat requests after my initial fumbling.

Sure, and that's an important tool for improving quality in the face of different skill levels / levels of effort. But that's independent of investing a bit in improving the skill level across the helper base; as you experienced, it's something where experience helps, and if your first time in the real flow is with an actual request, there's no way to avoid the helper needing to deal with both helping and learning simultaneously.

And that's just one more thing that blind people don't want to have to deal with when they just want to go about their lives.

Humans are slow and unreliable. AI is fast and consistent.

Both are imperfect. You shouldn't rely on either one for something life-or-death. Nobody is using them to decide if it's safe to cross a street.

Hydrogen is very difficult to use as a direct fuel source. This recent video gives a good overview of all the challenges a hydrogen engine has to overcome: https://youtu.be/DGL5g91KwLA

thanks

I agree that it's a great book; but I wouldn't recommend it as an entry point into understanding purely functional data structures.

Okasaki fleshes out a couple of examples and explains his thought processes and heuristics for developing such data structures quite well; but they're very much still notes on the early-stage exploration of the concept.

From a historical perspective it's still a fascinating read though and definitely recommend it if you want to read up on the origins of immutable data structures.

As a long time RSI-sufferer that has tried everything I can confirm the KeyboardIO was also very nice... but nothing beats Kinesis Advantage. I think it's the domed arrangement as I did feel the thumb arrangement on the KeyboardIO was superior.

I've bought a Kinesis Advantage about every three years since 2003 (I think). Still the best keyboard out there despite being USB-A and the lack of bluetooth. I really want to like the Kinesis 360, but the lack of a function key row keeps me from buying one.

I agree that I'd like to see more LiFePO4 battery packs; but compared to an EV I'd much rather have this happen on an e-bike/monowheel I can just step off.

Agreed, I was more thinking about it happening in the apartment. I do not see me surviving such event, unleast the battery is being charged on the balcony (but even then). LiFePO4 failure, while not a picnic, at least can be proofed against.

Of course, it is the same with scooters and bikes, but there are a lot of bikes (and almost no scooters) with removable packs, so yeah. But compared to the unicycle, bike takes up half of a small apartment.