[simonmesmith]

For sure. But would you really be a year behind? I don’t think it’s necessarily so. For example, if GPT-5 can write a full, deployable application from a text prompt with no errors, is it worth trying to hack something like AutoGPT to do it while waiting? It feels like in some cases you might be ahead and in others you’ll have wasted effort.

[warrenm]

of course you are a "year behind"

...because a year will have elapsed between now and a year from now :)

I started programming in the early 90s around age 10

And I was 10 years behind everyone who started in the early 80s. And 20 years behind everyone who started in the early 70s.

...and 10 years ahead of everyone who started in the early 00s

That's just math :)

If ChatGPT 5 comes out in a year, and you wait to start til it's here, you will be behind however long the developers of ChatGPT 5 have been working on it

AND you will NOT have a year of experience on ChatGPT 4 while waiting for version 5 :)

[simonmesmith]

But if you anticipate that GPT-5 will have X capabilities that will negate the need for you to learn Y intermediate tools, so you can instead focus on Z, you will have avoided wasting effort that could be better spent elsewhere.

Like, if I know we’re in a deflationary economic environment, I’m better off sitting on my cash and waiting for prices to drop before making a purchase, unless the benefit of owning that item outweighs the cost of not waiting.

I think I now do similar cost/benefit calculations regarding technology.

[syndicatedjelly]

> But if you anticipate that GPT-5 will have X capabilities that will negate the need for you to learn Y intermediate tools, so you can instead focus on Z, you will have avoided wasting effort that could be better spent elsewhere.

That's the gamble you play. It's pretty rare for a technology to have absolutely 0 transferable skills to the next one.