←back to thread

95% of Companies See 'Zero Return' on $30B Generative AI Spend

(thedailyadda.com)
418 points speckx | 1 comments | 21 Aug 25 15:36 UTC | HN request time: 0.207s | source
Show context
JCM9 ◴[21 Aug 25 16:22 UTC] No.44974621[source]▶
We are entering the “Trough of disillusionment.” These hype cycles are very predictable. GPT-5 being panned as a disappointment after endless hype may go down as GenAI’s “jump the shark” moment.

It’s all fun and games until the bean counters start asking for evidence of return on investment. GenAI folks better buckle up. Bumps ahead. The smart folks are already quietly preparing for a shift to ride the next hype wave up while others ride this train to the trough’s bottom.

Cue a bunch of increasingly desperate puff PR trying to show this stuff returns value.

replies(12): >>44974672 #>>44974721 #>>44974794 #>>44974830 #>>44974852 #>>44975105 #>>44975342 #>>44975381 #>>44975547 #>>44975669 #>>44975753 #>>44981385 #
1. pseudosavant ◴[21 Aug 25 17:35 UTC] No.44975669[source]▶
When OpenAI went from GPT-3.5-Turbo to GPT-4 it seemed massive, and there were no other steps in-between. And nobody else had a meaningful competitive model out yet.

When GPT-5 came out, it wasn't going from GPT-4 to GPT-5. Since GPT-4 there has been: 4o, o1, o3, o3-mini, o4-mini, o4-mini-high, GPT-4.1, and GPT-4.5. And many other models (Llama, DeepSeek, Gemini, etc) from competitors have been released too.

We'll probably never experience a GPT-3.5 to GPT-4 jump again. If GPT-5 was the first reasoning model, it would have seemed like that kind of jump, but it wasn't the first of anything. It is trying to unify all of the kinds of models OpenAI has offered, into one model family.