Ploum

But they aren't useful for that. Terry Tao uses them to improve his ability to use poorly-documented boilerplatey things like Lean and matplotlib, but receiving advice from them‽ Frankly, if a chatbot is giving you much better advice than a rubber duck, you're either a Jack-of-all-Trades (in which case, I'd recommend better tools) or a https://ploum.net/2024-12-23-julius-en.html Julius (in which case, I'd recommend staying away from anything important).

I recommend reading his interview with Matteo Wong, where he proposes the opposite: https://www.theatlantic.com/technology/archive/2024/10/terence-tao-ai-interview/680153/

> With o1, you can kind of do this. I gave it a problem I knew how to solve, and I tried to guide the model. First I gave it a hint, and it ignored the hint and did something else, which didn’t work. When I explained this, it apologized and said, “Okay, I’ll do it your way.” And then it carried out my instructions reasonably well, and then it got stuck again, and I had to correct it again. The model never figured out the most clever steps. It could do all the routine things, but it was very unimaginative.

I agree with his overall vision, but transformer-based chatbots will not be the AI algorithm that supports it. Highly-automated proof assistants like Isabelle's Sledgehammer are closer (and even those are really, really crude, compared to what we could have).

Sounds already too complex for me.

I wrote my post as Gemtext (the Gemini markdown) and wrote a python script that convert it to HTML and generate an index page.

Also, I put the CSS in each page because I’ve 42 lines of CSS and I did the math to show that having 42 more lines in each file would never cost more than having a second http request.

Honestly, it was a lot simpler than expected, without using any python library or any template. The hardest part was generating the email to send to the list. I spent month analysing different frameworks, different static site generators. I spent week trying to find a good CSS.

According to the git history, it took me 2 weeks of hacking to get it done including importing 18 years of Wordpress history. And without having to learn anything because it was straight python. And the best of all is that I can easily change/correct if I’ve a bug. It’s my code.

We, computer scientists, forgot completely that we could simply code adhoc solutions instead of trying to reuse everything. We take more time to learn and adapt than to do.

https://git.sr.ht/~lioploum/ploum.net for the code

(the site itself is https://ploum.net)

> great companies can have an enormous upside for countries (e.g., taxes and employment)

A company split into multiple still provides taxes and employment. And the employment conditions tend to improve, since there would be more companies competing for employees.

> too strict actions may hamper innovation by signalling to innovators that you will be punished if you achieve too much

And monopolies hamper innovation by reducing the competitive pressure to innovate. Necessity is the mother of invention.

Let's look at what kind of "innovations" Microsoft used to grow - pushed an over-complicated OOXML standard, and implemented it in their products incorrectly, to make it harder for LibreOffice to interoperate, since they were forced to follow a complex standard and reproduce Microsoft's bugs: https://ploum.net/2023-06-23-how-to-kill-decentralised-networks.html

There are countless similar examples, going back to railroads charging more to carry competitor's goods.

Also, keep in mind that patents don't mean the invention works, and they certainly don't mean the invention is useful for anything.

https://ploum.net/working-with-patents/

> Note that by « valid », I mean that the Patent Office didn’t found a trivial prior art for this. It doesn’t mean that there is no prior art or that I’m the real inventor or that my invention works.

An example of an insane patent:

https://www.google.com/patents/US6960975

> Space vehicle propelled by the pressure of inflationary vacuum state US 6960975 B1

(via https://www.metabunk.org/do-patents-mean-the-invention-works.t194/ )

> Otherwise I can spread all kinds of made-up stories about my work mates, and claim they illustrate truths about human nature, right?

This happens all the time, and it’s fine? A recent front page example [0] — did julius ever exist? Who knows. Would it change anything? Not really…

A problem could definitely arise if you specifically called out a coworker without anonymization, but speaking in broad strokes is… perfectly fine and uneventful

[0] https://ploum.net/2024-12-23-julius-en.html