MattHeard

it can tell you if your spec is bad, but it can't tell you if your spec is good

One thing I haven't seen addressed in the post or comments (sorry if it's there and I just missed it) is for apps and websites that sell non-digital goods to customers.

For returning customers, a customer with an app installed is simply going to have a lot less friction to open the installed app and place an order than a customer who has to open their browser, log in again, and then place the order.

That alone is worth the investment for many companies in making an app, even if a minority of customers actually choose to install the app and keep it installed.

> It doesn't have to be actively not fun either, but the primary feeling should be that of effort.

Karpathy clearly wrote something different that what you say he wrote.

I use Github Copilot with PlantUML. I use PlantUML comments to describe what I want the diagram components to show and then the syntax is written for me. I can then edit it pretty easily and iteratively.

It appeared for me about thirty minutes after I first checked.

Hello, thank you for sharing your thoughts on this topic. I'm currently writing a blog post where the thesis is that the root disagreement between "AI doomers" and others is actually primarily a disagreement about materialism, and I've been looking for evidence of this disagreement in the wild. Thanks for sharing your opinion.

I got as far as here:

> If you need to introduce a new pattern, consider refactoring the entire codebase to use this new pattern, rather than creating one-off inconsistencies.

Putting aside the mis-application of "pattern" (which _should_ be used with respect to a specific design problem, per the Gang of Four), this suggestion to "refactor the entire codebase" is impractical and calcifying.

Consistency increases legibility, but only to a certain point. If the problems that your software is trying to solve drift (as they always do with successful software), the solutions that your software employs must also shift accordingly. You can do this gradually, experimenting with possible new solutions and implementations and patterns, as you get a feel for the new problems you are solving, or you can insist on "consistency" and then find yourself having to perform a Big Rewrite under unrealistic pressure.

the pythagorean theorem is not named after its discoverer, unfortunately, as historians of mathematics have dated its earliest known usage to well before pythagoras

If the quality is too low, it's not viable. This was always true.

i was going to ask if turtlebot was too small for you, but thought i should check the price first so uh, yeah, i'm guessing you're thinking the 250 price point instead of the turtlebot's 1000+

My working assumption has always been to check the docs of your specific regexp parser, and to write some tests (either automated or manually in a REPL) with specific patterns that you are interested in using.

"I would like website operators to assume that I consent to being tracked, so I'm annoyed that website operators are not allowed to assume that everybody consents to being tracked."

surprisingly common to find oneself caring strongly about bullshit

I have/had this cat water fountain too (my cats are now happy with a normal bowl) and I believe I just threw the wall wart into a box with the rest of my unused USB A wall warts. A little bit like electrical Russian roulette now, I guess.

Barely related, but I remember the joy as a kid of putting a computer game CD-ROM in my Walkman CD player just to see what happens and discovering that I could listen to the game's soundtrack just like a normal soundtrack CD. Can't remember which game, but it is a fond memory of mixed-mode CD technology. Can't do that with your Steam games and Spotify!

I was hopeful for a private-industry approach to AI safety, but it looks unlikely now, and due to the slow pace of state investment in public AI R&D, all approaches to AI safety look unlikely now.

Safety research on toy models will continue to provide developments, but the industry expectation appears to be that emergent properties puts a low ceiling on what can be learned about safety without researching on cutting edge models.

Altman touted the governance structure of OpenAI as a mechanism for ensuring the organisation's prioritisation of safety, but the reports of internal reallocation away from safety towards keeping ChatGPT running under load concern me. Now the board has demonstrated that it was technically capable but insufficiently powerful to keep these interests in line, it seems unclear how any safety-oriented organisation, including Anthropic, could avoid the accelerationist influence of funders.

I've heard the opposite about Brockman. What makes you so confident about this tech abilities?

Maybe this news should challenge your priors, then?

A lot of courts use rulings from other countries where the cases are similar. Any interpretation by the Berlin regional court that GDPR implies that DNT should be treated as a GDPR opt-out should be easily adopted by courts in other countries deciding similar cases.

The licence is not for the user, the licence is for the publisher to do the bare minimum needed to tell their bosses that the model weights or implementation code was published ethically.

As others have pointed out, these licences appear unenforceable.

The publisher simply wants to appear responsible. There are likely many open-source-oriented engineers and scientists at these tech firms who have been pushing for publication. (See the discussion between Mark Zuckerberg and Lex Fridman.) The involved tech firms only care about the licence as far as it might minimise the likelihood of public backlash if any of these published models cause harm.

"If you're physically capable of coming in 5 times over the course of two months you must therefore also be physically capable of coming in 26 times over the course of two months."

(Sarcasm, of course.)

I prefer to stop at 80%.

some consumers will pay the premium of ownership over the licence and some won't. if either market has demand fall below a particular threshold, the market will collapse. this isn't a moral issue, this is simply an effect of a new market for a service substituting effectively for an old market for a good. it's clear that a lot of previous book-owners really just wanted to be able to access a text

if there is really a mass misunderstanding by ebook consumers that they don't have rights as part of their licence that they believe they have, the market demand will start shifting back towards ownership

When I see this criticism of mocks, it appears to me that the critic is primarily used to seeing tests that are either mocking the thing-under-test or don't understand the boundaries of the thing-under-test.

I used to be an active user of all the old Google Labs tools but now my personal Workspace account locks me out of everything, so meh.

Germany's electricity generation from coal has over the long run substantially decreased. The low of 2020 appears to correlate with the COVID-19 pandemic and the increase since then appears to be more of a reversion to the original trend than anything to do with changes in electricity from nuclear.

2022 coal use is a third greater than 2020, as _dain_ mentions, but 2020 was anomalously low usage rather than representing a normal baseline for comparison.

While any increase in coal usage is a problem, the coal use should be compared to the average of the last few years with the anomaly of 2020 removed. This would produce effectively no increase or decrease in coal usage.

Source: https://ag-energiebilanzen.de/wp-content/uploads/2023/03/STR... via https://ag-energiebilanzen.de/

Are server-side changes needed?

GDPR fines are intended to generally be proportional to the company's revenue.

This is exactly the kind of advice I give to people who ask me about teaching themselves to "code".

1. Find a thing you want to make. 2. Find out how to make it. 3. Try to make it. 4. Learn the skills that previously prevented you from making it.

Both events are extreme yet possible.