https://bsky.app/profile/iwriteok.bsky.social/post/3lwcvfzwjuc23
Robert Evans quoteskeeting dgerard's mic drop
this post was submitted on 11 Aug 2025
22 points (100.0% liked)
TechTakes
2311 readers
98 users here now
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
founded 2 years ago
MODERATORS
Can anyone explain to me why tf do promptfondlers hate GPT5 in non-crazy terms? Actually I have a whole list of questions related to this, I feel like I completely lost any connection to this discourse at this point:
- Is GPT5 "worse" in any sensible definition of the word? I've long complained that there is no good scientific metric to grade those on but like, it can count 'r's in "strawberry" so I thought it's supposed to be nominally better?
- Why doesn't OpenAI simply allow users to use the old model (4o I think?) It sounds like the simplest thing to do.
- Do we know if OpenAI actually changed something? Is the model different in any interesting way?
- Bonus question: what the fuck is wrong with OpenAI's naming scheme? 4, then 4o? And there's also o4 that's something else??
- from what i can tell people who roleplayed bf/gf with the idiot box aka grew parasocial relationship with idiot box did that on 4o, and now they can't make it work on 5 so they got big mad
- i think it's only if they pay up 200$/mo, previously it was probably available at lower tiers
- yeah they might have found a way to blow money faster somehow https://www.tomshardware.com/tech-industry/artificial-intelligence/chatgpt-5-power-consumption-could-be-as-much-as-eight-times-higher-than-gpt-4-research-institute-estimates-medium-sized-gpt-5-response-can-consume-up-to-40-watt-hours-of-electricity ed zitron says also that while some of prompt could be cached previously it looks like it can't be done now because there's fresh new thing that chooses model for user, while some of these new models are supposedly even heavier. even that openai intention seemed to be compute savings, because some of that load presumably was to be dealt with using smaller models
load more comments
(5 replies)
This was discussed last week but I looked at the comments and noticed someone in the comments getting slammed for... checks notes... noting that Eliezer wasn't clear on what research paper he was actually responding to (multiple other comments are kind of confused, because they assume he means one paper then other comments correct them that he obviously meant another). The commenter of course edits to back-peddle.
load more comments
(3 replies)
"usecase" is a cursed term. It's an inverted fnord that lets the reader know that whatever follows can be safely ignored.
If I ever get the urge to start a website for creatives to sell their media, please slap me in the face and remind me it will absolutely not be worth it.
Palantir's public relations team explains how it helped America win the Global War on Terror