whiplash451
whiplash451 on Hacker News
My bad I stand corrected.
I think open weights is a fair term, much like open source but for weights.
Open source never implied that commercial applications were authorized.
I feel like the community is splitting hairs when the Anthropic/OpenAI are just locking everybody in way further than Mistral.
Not just more dimensions, but off-the-shelf multimodal encoders
Can you please share:
- what specific input you see this issue with?
- which interface you used (API, playground, specific sdk version or a wrapper app)?
- whether the chinese chars appear in the thinking section or the output?
Thanks.
To be fair, they let you build a $200M yearly revenue company on top of it for free, seems quite generous
> it's hard not to look at it as marketing
that does not sound fair given the depth of the blog post, at some point we need to figure out what is marketing or not ourselves
There's serious arbitrage available for a startup to come in and help with this mess. Like: scan your mouth with your iphone and get a recommendation. I'll be called crazy, but ...
I see correlation instead. You IPO when you grow like crazy. You build something like scion when you grow like crazy.
> A team of seven people built both Scion and Antique (our code review tool, which we could easily write a whole other blog post about) [...]. We did all this while the volume of code merged per day increased by 10x
Truly impressive, given the scale of Anthropic. Speaks a lot about the level of talent there, but also how far/fast we're moving into a new world.
> Practically speaking, it’s difficult to offer Scion in a form that would be broadly useful to other organizations.
Alas... open-sourcing scion would have been the true generous move.
And? 5-7 years is nothing at the scale of our modern society and the outcome of what you describe could be enormous.
OTOH "$105 million to Claude Code over a 28-day timeframe"
$1.4B/year is not a small number, even at Meta's scale, when it's money going to a competitor
For inference (Anthropic and OAI are B2C on top of B2B)
To train much larger models. It is quite possible that 10T-100T models be on the horizon
They don't use public money to produce them (and maintain the museum).
It is self-funded. Public, but self-funded.
So this has nothing to do with misspent public money.
The Rodin Museum is a piece of jewel in Paris that exists only because Rodin himself decided for things to be run that way on his deathbed.
The wish of the artist overrules the wishes of random people that "consume" his work.
The French law considers the first 12 bronze productions of a clay cast as an original.
Article R. 122-3 of the Code de la propriété intellectuelle
I think you’re confusing fan in and fan out.
As an engineer, I can relate but also found a workaround in deciding what kind of interactions I wanted to have with an LLM. As a coding agent? Yes. As an advisor for life decisions? No. And then there's a lot of in-between where I decide to use an LLM or not based on whether I will dislike the tone.
OpenAI and Anthropic are so close on benchmarks and release date that they are essentially ex-aequo at this point. The market is just hedging their bets.
There is no close second, because the AA Index points are expentially harder to get as you get closer to the first.
There’s quite a few companies in that field already. Keen to hear their feedback if they are reading this.
Is there evidence that small models hallucinate less than big ones?
Groundbreaking includes the “explain to the world why it matters” part
Well I surely hope so. That’s the whole point of fine tuning.
Why inefficient? If Claude continuously makes the system more efficient based on demand, there's no reason for the system to be less efficient than if it was managed by humans.
I am with you and wish you were right, but good luck forcing Meta to change the key dark design patterns of their products (correctly identified by the regulators as "highly personalised recommendations, autoplay and infinite scroll")
This is a step in the right direction, though. It will be a long journey.
You pay your taxes to Lightspeed?
dang does not scale. It’s on us to make this place how we want it to be. Every comment matters
> negative feedback - and dismissive feedback - is something you need to be prepared
OK. Let's RFM: first line = "Be kind. Don't be snarky. Converse curiously; don't cross-examine. Edit out swipes."
So, no, it's not HN-OK for rude & unfounded feedback to make it to the top comment.
Even if the website is "bad" (i.e. those 30 papers were never linked to Ilya etc.), the proper way of handling this is to call it out in a neutral way.
We might be over complicating things here.
The governments’ focus might be on protecting genuine users (adults or not), not fighting fraudsters.
In other words if ZKP works for the vast majority of technically illiterate people with their EU ewallet, the job is done.
That sounds like a lot of complexity for the average user.
Aren’t there fees to convert fiat from/to crypto on both ends?