AI Crap

When I read things like this it makes me think the author hasn't used ChatGPT in their job yet.

Here is a really simple example of how I used ChatGPT this afternoon that saved me, I would estimate, about 2 hours of work:

I had 2 CSV files, with different formats but which (supposedly) had the same functional information in them.

I had a very complicated BigQuery SQL statement that worked on the first file format by importing it as a blob to a table then combining it with a bunch of other CSV files. I wanted to know how much I might need to change my query if I started using the 2nd CSV file (which takes much less time to export from the system that produces it).

The query of course has a big complicated SELECT statement, but also several common table expressions and joins, some of which use columns from the CSV file I was looking to replace.

So I gave ChatGPT the 2 header rows, and the big complicated query. I asked it to tell me likely mapping between the 2 CSV files for similar columns, and to give me a list of columns that appeared in one but not the other. I asked it to mark with an exclamation mark those columns which appeared in the query.

It got some of the things wrong but because I'm pretty familiar with the query and the files I was able to pick up on those and it was much, much easier to browse the output and pick out the errors than it was to break down the query and do all that analysis from scratch.

The whole process using ChatGPT took me about 15 minutes.

And I have wins like that I would say about once per day. I mean it: I'm saving probably about 2 hours work per day by using ChatGPT on average, on tasks just like this.

Now multiply that by all the shit that people are doing all the time and think about all the needs that will get met as a result of this increase in productivity that are not currently being met, and you have an idea of why AI is fucking awesome, ESPECIALLY given the fact that we need a decreasing working population to support an increasing retired population.

Every (tech) sword is triple-edged, and it's the third one that people trip up on because they never expect it.

That code generation is "banal" to you, or a threat to junior developers, suggests to me that it's already huge.

Likewise the job replacements you fear: that's only even possible when the AI doesn't suck massively.

The current state of downloadable models implies any corporate advantage is temporary.

We've seen time and time again that "the people" prefers centralised, paid, convenient management of complexity.

If the majority of "the people" prefers paying a subscription for watching movies or listening to music, I doubt they'll make the effort to learn how to tune and run private LLMs locally, for medical aspects or otherwise. Not when there will be major companies spending billions on marketing more convenient options.

Guillermo Rauch's essay [1] still rings true: it's hard to forego efficiency (though it works just as well for convenience).

[1] https://rauchg.com/2017/its-hard-to-forego-efficiency