@kiwitalkz 1B on marketing seems unrealistic. even Endgame only did 150-200M on marketing, i doubt TTWO needs to spend anywhere near that much given the game hypes itself even among casual audiences. it'll probably be a 100-200M campaign
@deredleritt3r@danshipper what about xhigh? according to deep-swe and other benchmarks, the max effort has ~no gain over xhigh, so it seems like a waste
@randomdude22401@synthwavedd@kimmonismus the only model they've released on friday was o3-mini. dont get ur hopes up
for now im believing polymarket which indicates next week prolly june 11th
@argofowl i remember people said use up all your usage the day before 5.5 came out because "they'll reset it"
... they never did
don't make people panicwaste their usage ffs
@JinjingLiang > The Codex app actually looks good
it's looked good for ages because humans are steering it...
GPT-5.x has been "good" at UI for ages if you have someone who knows what they're doing steering it
only useless vibe coders who can't design for shit think GPT is bad
@joshbuildings@hsienshoryu@deredleritt3r already all over X, people are saying "ChatGPT reached 1B MAU!!!" including literally prinz himself
you're saying:
web + mobile: 900M *W*AU still
mobile only: 1B *M*AU, new milestone
just very confusing and no one layperson will know the difference here
@joshbuildings@hsienshoryu@deredleritt3r ugh idk the reporting is very ambiguous
if you say "app" you need to say "mobile" with it to be specific that it's limited to iOS + Android, because "web apps" are a thing and what ChatGPT is
@hsienshoryu@deredleritt3r nvm this is just according to some slop analyst firm called Sensor Tower, not from oai themselves
i guarantee they've been at 1B MAU for months, since last year around Q4 at latest
@hsienshoryu@deredleritt3r no... it's definitely web + mobile combined, and their only other product is codex w 5m users and many of those probably already use chatgpt
you would think 900M weekly uniques was already well over 1B monthly uniques, but... it wasn't
@OrenMe literally no one is using this in the big june 26 when codex and claude offer actual practical limits
github copilot should pack it up and move on, it will never be relevant now
@GWHayduke97 but it got 3.5K likes and he's still at his shtick to a degree (pivoting to financials rather than capabilities) so we can continue to trust his insights <3
@haider1 early April when Anthropic announced Mythos was the last we felt a step change (5.5 wasn't over 5.4 imo)
people become accustomed to the new norm extremely fast now, so 2 months is ages
and it doesn't help that all the new Opus releases are mid af, so progress has felt slower
@edzitron these are not the "actual costs"
the labs are forcing microslop to pay full api prices which is not the real compute costs (which are 5-10x cheaper)
so using $2000 in api credits is more like $200-$400 in real costs which is why first party subs work unit wise
@alexecutedev@jasperdevs true but at worst its an 8-12 hour differential unless you dont use it after it resets. so close enough; at least a million devs are probably within 10 min of each other
@koltregaskes this is the main reason humans remain in control for now, because agents need handholding for all their goals.
and if oai can't fix this, then they'd be true to their recent messaging pivot that humans won't be 'replaced', but that means agi isn't possible. so i doubt it.