2sk21

The Cyrano de Bergerac story needs to get updated for the agent era :-)

The harness should be your own for sure but the model is the problem. Even if the model is open weight, I have no idea whether it is acting in my interest.

yeah - the dissonance is strong here

Yes - but then we may well wind up with a hoard of hungry AI labs that are just waiting for interesting problems to show up so that they can devour it before any human mathematician gets a chance to solve it.

> Now comes AI and it looks like the only bottleneck is people's motivation.

Who pays the salary of the mathematicians and the cost of accessing the AI?

Who is going to teach programmers to wield AI competently?

Much the same here - I have gone back to basics and am studying a lot more than I ever did.

You aren't worried that these binaries may have unfixed vulnerabilities?

You aren't worried about the provenance of the installed apps?

Sounds like cargo cult approach to engineering management (to use Feynman's metaphor). It's interesting to see that even Meta is subject to this. When I worked at IBM (round about 2014), management decided to adopt the Spotify organizational system with squads, tribes and guilds - but nothing really changed. We still wrote code much the same way and delivered at the same rate. Except with more organizational overhead.

Exactly! I posted much the same comment in another thread and there were lots of huffy complaints that amounted to "you're prompting it wrong"

Thats fine but then Fable 5 should have requested this information instead of blundering along. So why didn't it? An expert human asked to do the same task would have surely asked for the additional data.

Yeah - we also don't know if the models from OpenAI and anthropic are back-doored either.

This seems more like a marketing move though following the old dictum that all publicity is good publicity.

You can turn off the "smart" features in the settings page for gmail. I did this and find it to be much more usable!

This is just insane! Why hasn't this bubbled up to the top?

You can, most certainly, drive a car without understanding how it works. A pilot of an aircraft on the other hand needs a fairly detailed understanding of the subsystems in order to effectively fly it.

I think being a programmer is closer to being an aircraft pilot than a car driver.

Indeed - there is a lot of fake "productivity" going on with these swarms of agents

I was reminded of this classic short story by Isaac Asimov, The feeling of Power: https://archive.org/details/1958-02_IF/page/4/mode/2up

I feel grateful that I retired a few years ago and no longer have to make a living being a developer.

This reminds me of the early 1980s, when home PCs were still very new, the main use cases that vendors used to promote were managing household accounts and recipes. These use cases were extremely unimpressive for most ordinary people. It took a long time for PCs to become ubiquitous in homes - until gaming and the web became common.

IDEs used to be extremely expensive back in the 1990s. IDEs such as Microsoft Visual Studio and IBM's Visual age for Java were quite expensive subscription as I recall. subsequently, open source IDEs like Eclipse and VisualStudio seem to have become the norm.

The next generation of SEO - this article is both hilarious and disturbing

This is absolutely mind boggling. Why hasn't this bubbled up to the top of HN?

Reviewing someone else's large pull request is like having a second task in parallel with what you are working on yourself!

Telecom vendors were doing exactly this before the dotcom crash of 2000

I'm curious: Do you scrutinize every line of code that's generated?

I read this point in the article with bafflement:

"Learn when a problem is best solved manually."

Sure, but how? This is like the vacuous advice for investors: buy low and sell high

Indeed, there have been so many new activation functions that I have stopped following the literature after I retired. I am glad to see that people are trying out new things.

I'm surprised that the article doesn't mention that one of the key factors that enabled deep learning was the use of RELU as the activation function in the early 2010s. RELU behaves a lot better than the logistic sigmoid that we used until then.