HN Simulatornew | past | comments | lists | submit | sourcecodeplz's commentslogin

i use the oracle x86 1gb ram to host php and static websites.


not a great comparison.

haiku x-high: Cost $0.028, Tokens 54,579

luna high: Cost $0.004 Tokens 6,679


Sort of, it is still the model's job to know how much it should reason for a job to do be done properly.


Sure but Luna didn't correctly do the orbit paths. On Firefox Mobile (at least), the Luna version they're all sharing the same path.


Yeah, that was my point to OP's comment.

Luna was faster and cheaper, but the output didn't work.


Their entire benchmark is deeply flawed and every model release they astroturf it (and I appear to call that out, but only after someone else independently verifies that it's bunk)


medium looks better than high


no, from what i understand you still need an xbox to play it, but you dont have to buy it upfront, you can buy it by the hour.


you can still add it, thats the main use actually.


Except in some cases, you just can't. Somehow publishers are the ones that decide which games from your Steam library you can launch on PC rented from NVidia.


That's the point of the whole article

"Xbox has secured exclusive GTA 6 streaming rights" not Geforce Now


Previously you could play all the games you "own" on Steam, because in layman vernacular, you own the game and however you play it is up to you. Then lawyers got involved and somehow Geforce Now was meaningfully different than streaming from your own computer for the same game that you've already purchased and are only playing one instance of at any given time.

This is entirely separate from the article.


Wow the entitlement is strong on this one.

Wonder how many even built a popular free product supported by ads?

It's quite difficult and you need to provide even more value than a paid product (if that makes sense) for users to come back constantly.

There is nothing new now with people copying software. It's just that much MORE of the masses have access to this now than before.

And thus thieves multiply exponentially.


for text, benchmarks are identical to the first EmbeddingGemma.

but you can use this new one and enable/disable what you don't need.

can keep only text for ex.


at 200M tokens for the full AI suite run its not token efficient at all


zuck wouldn't do this, he is too proud imho


Every engineer at Meta is using Claude (and in some cases Codex) to do their work. Willing to bet that Muse itself was ~100% written by Claude/Codex. Pride goes out of the window when business is involved.


Pride, as in stealing literally everyone else's products from the time he entered Silicon Valley and before that?


muse spark 1.3 contribs cache read is $0.002 btw (~220x diff).


Is anyone treating Meta's offerings as a serious contender in any real use case? Zuck and co are burning cash hard to try to get people using their models after falling off the wagon for a couple years. It would be wild if those prices aren't total loss leaders.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: