[Juliate]

Guys guys. I do share the sentiment.

On the other hand. Unless we have a breakthrough hardware/physical innovation, GenAI in its current forecast is not energy-efficient, cost-efficient, compared to human/deterministic methods. It has shown no capacity so far to « create » in the sense we animals do.

And all that, still being: highly subsidized (your subscription does NOT cover the costs of the service as of now, we are still in the market creation/capture phase), and without mesurable economical benefit.

Things are still far from being over.

[orangecoffee]

To give you the benefit of doubt - you're not paying attention. The 100$ claude account is profitable inference wise, and performs better than 80% of engineers with upto 5YOE. Architects I speak with want to delegate to this 100$ agent rather than 10000$ engineer.

It is absolutely more cost efficient. Junior engineers are done for.

[Juliate]

I have first hand and second hand seen more failure from this line of reasoning _as of today_, in the French/European tech sector, than successes. Not even talking about those executives that expect to use Gen AI to "help" in their process and financial reporting - at least the flag is bright red.

I've also met enough architects (or so called) in consulting, service companies as well as startups, from both a founder and a tech lead point of view, in the past 25 years not to entrust them too much with their hunches, long-term wise.

Especially because, from a sustainability point of view, saying "junior engineers are done for" is like saying "kids are done for": not that wise. Unless what you account/bet for is the drastic reduction/disappearance of humains (and then, a lot starts to make more sense).