if its that simple, why doesn’t Anthropic just distill its own models and release Fable 5.1, 5.2, ...?
HN user
laweijfmvo
ditto, it’s just not stable or usable at all for me.
agree, not being able to guarantee or even know, it seems, how your sizing works is a deal breaker and i might have to cancel my order until i see how the first reviews come out :/
After a few minutes on the site, I have no clue what this is for. A keyboard that interacts with Codex? That’s just a software feature, why am I paying $230 for hotkeys?
1. Starlink, which you provided
2. Made the modern EV relatively commonplace; no other manufacturer was taking it seriously until Tesla succeeded, and took many years to catch up, although they have
3. Re-usable rockets / higher launch cadence leading to significantly cheaper costs to put things in space. No major competition yet.
I discovered and tried to use this feature to turn an older iPhone into a dumb phone for myself, but hit several blockers
1. It’s incredibly slow to transition in and out of the mode, as mentioned in the article, which made setting it up (constant tweaks) very painful
2. For messages and calls, you were limited to select contacts only. So I couldn’t just text/call a number when I needed to.
They may have polished the feature since then, but given that it’s an Accessibility feature and was never meant to receive much attention in this regard, it may always be half-baked.our wages are frankly astronomical compared to my parents’, and we’re DINKs and only one of my parents even worked. Guess which one of us owns a house.
I think Meta’s massive compute investment was never about its 100,000 engineers running coding models, but its 3,500,000,000 users wanting to use AI in every single product (and some new ones: Meta AI, glasses, etc.) So I would think that’s the part that’s not being utilized anywhere near the amount they hoped...
for us, maybe, but for someone who never really used the workflow, or looked at the “thinking” output where models spin their tokens on the stupidest shit, i can see how it wasn’t obvious.
in fact i’d say it’s a rather poor wrapper and sensationalizes far beyond what reuters actually reported. e.g. the unsupported claim that Zuckerberg somehow put Alex Wang “in charge of the entire company.”
his biggest fear is not being the one to nail the next big thing. it happened with mobile, although that was basically irrelevant in hindsight. he thought it was metaverse, which failed, while in the meantime it was actually the new AI cycle, which he was late on / lost by the time llama 4 flopped.
i think he’d rather just blow up the whole company than continue to be solely associated with one (or even a few) of the most successful websites/apps in history, for some reason. maybe he thinks people will like him if he does something else?
huh, but what if the AI trashes my git repo? maybe it just deletes the .git folder entirely. a deterministic undo wouldn’t be the silliest feature, for the current definition of “AI”.
> Completely trained from scratch new models will be released by @SpaceX every month this year.
doubtful…i think the article is saying that’s what the book claims, not whether it’s true or false.
it would be one thing if congress passed a bunch of (probably inadequate) legislature that every AI company had to follow to operate in the US. it's another when it's a faceless/nameless group of people probably deciding arbitrarily based on mood and bribes.
are the cops gonna do anything? just start a mini civil war in someone’s garage?
I don’t read the Minnesota Star and hadn’t seen any of this, so, if anything the Feds doxxed their own by bringing it to my attention.
i'm still trying to figure out why the Eiffel tower is featured in the first graph
and it will look terrible if you don’t do it properly
i don’t really understand the obsession with 10g ethernet, to the point of complaining about only getting 8g performance. if you need those last 2g so badly clearly you should be investing in proper network hardware and not some dinky piece of plastic connecting over USB?
if only there was a place to share them, and that place actually showed them to the people who indicated that they wanted to see them.
i assume they also owe back payments? with interest?
His answer was “or what?”. And then enough people they cared about quit, and now you’re seeing the articles about them apologizing, saying they care, etc.
So the petition didn’t work, exactly, but served as a warning: employees said, fix this (or else). Mark said “or else what?”, and enough people said “I quit” to make them look bad.
also NAL and i think you’re right, the petition even says as much. but at-will employment means they can fire you for basically any other reason, and they probably have at least 10 to choose from at any given time.
it’s worth paying attention to what the stock price was 4 years ago for the next year or so
but who’s going to pay Meta for it?
A Waymo ride once quoted me close to $100 for a couple miles. The same Uber ride ended up being like $21. Surge pricing is real but c’mon.
if your boss suddenly asks you to carry a pager, do not accept it.
so how long was the bug there? was there a way to access it before/without the support agent? it feels like Meta will throw anything under the bus to redirect blame from the AI, because that would be the end of their $600B (depending on “which number you want to go with”) experiment
> The industry is much closer to self-improving AI than previously expected, they warn.
Uh, no. Maybe we’re getting close to AI-assisted AI research, but nothing is going to happen unless a human presses the buttons. > So tech companies should consider slowing or pausing frontier AI development to allow time for researchers to gain a better understanding of the potential societal harms self-improving AI could unleash.
Ah, so they’re stuck and they want to slow their competitors down while they think of something else.