TechTakes

It never ceases to amaze me watching these tech bros breathlessly hyping the next revolutionary product for it to be a huge nothing burger. Like that's it? You gave gpt an image to text input + and a text to speech app? Dawg, I've been watching Joe Biden & Trump play Fortnite for 5 years at this point, text to speech just doesn't hit the same anymore.

I mean my god, at least some hardcore engineering went in to making the migraine inducing apple vision pro. If this is the best card in oai's deck, I would be panicking as an investor.

Well, I suppose dating a chatbot taped to a fleshlight is exactly the sort of relationship these people deserve. Self-solving problem.

It got submitted to lobste.rs too and there are some insane takes there too

https://lobste.rs/s/ph5o0a/openai_gpt_4o

this model has lower latency, and they’ve priced it at half of the previous model’s price, despite it having comparable or better performance. If this is not a dumping price,[...]

Yeah it's totes more efficient, it's not like Altman/MSFT are desperate to goose subscriber numbers

I’ve seen first-hand how it helps non-programmers write programs to solve their problems (with “bad” code, but bad code is still empowering people), and I think that’s important given that software has been eating the world.

Couldn’t the thousands of people now free from mundane work, work together now on solving climate issues or issues that require physical human labor?

if LLMs/AI can run byrocracies (and whole countries) more efficiently, there are potential energy savings there too.

ChatGPT-4-o-face

Talking about the demo and not those HN reactions (they suck): So we all in tech have collectively forgotten the whole 'apple iphones didn't actually work when they were being presented and it was all just theatre' thing? Because I wouldn't trust a slick tech demo like that at all. (I also want to know about all the edge cases where it fails and well, there is also the whole 'what is this even for?' thing). (And yes some HN people did voice these concerns).

E: also don't these GPT things have a sort of token limit? After a while they just get stuck in 'cannot accept more new data' mode basically (which I assume is a limit so people don't use up too much compute (I hate how we are making that a word now)), how would that work? (This is more about the robofuckers, who have also not learned anything from the previous gen of this shit when somebody turned the naughty cybersex function off).