usef-

9 ms·

181.
by usef- 4mo ago
You mean from the private investors? It seems the labs on both sides of the ocean are quite negative in their profitability right now due to the competitiveness. Though Anthropic claims they will have a profitable quarter this year (despite
182.
by usef- 4mo ago
Oh I know. But people often point to the API usage cost as an indicator of the magnitude of subsidisation, or to say that the big labs are far less efficient than cheap competitors. I'm saying that this is not necessarily the case. The
183.
by usef- 4mo ago
Z.ai is also believed to be "subsidised". Its parent company is running at a large loss right now. Anthropic have claimed they expect their first profitable quarter this year -- they may have bigger margins on their raw API than y
184.
by usef- 4mo ago
I believe grandparent meant "small enough changes that you can understand what the effects are likely to be"
185.
by usef- 4mo ago
Maybe someone else claimed it?
186.
by usef- 4mo ago
I think what he means is that the iPhone dots seem to move based on vehicle acceleration and deceleration, whereas the android ones use all phone motion. If I move my iPhone around in my hands the dots don't move, but on my android the
187.
by usef- 4mo ago
> What mitigations? They did try to place limits on their API, and tried to develop classifiers for AI-vs-non-AI text (which was abandoned in 2023, in a world of many models). A lot of their efforts in those days seemed to be to work wit
188.
by usef- 4mo ago
That line of thinking (public goods) is why the same people started OpenAI as a non profit originally. Notice how almost no Universities are producing large models? A key problem is that orgs can't get enough funds to stay on the front
189.
by usef- 4mo ago
I said this in the other thread, but they were proven right about their gpt2 worries, weren't they? From the original 2019 release: > We can also imagine the application of these models for malicious purposes , including the followi
190.
by usef- 4mo ago
What do you think they would do differently if they were genuinely worried about the safety?
191.
by usef- 4mo ago
To be fair, they were proven right about automated spam, phishing and disinformation being a problem. Yes, some of it looks silly now, though it's always easy to criticize with hindsight: the models could do unexpectedly impressive thi
192.
by usef- 4mo ago
They aren't saying there's a 100% chance of doom. They believe there's a non-zero chance of doom so would rather an org that prioritises safety to be the one at the frontier, on the assumption (I presume) that there will be a
193.
by usef- 4mo ago
Yes, I believe the reasoning is that they think safety research can best be done from the frontier. If you believe it will be developed regardless and that that there's a 30% chance of doom, they want a company prioritising safety rese
194.
by usef- 4mo ago
This policy change doesn't allow training, just like the previous one.
195.
by usef- 4mo ago
Note that the terms still prevent them training on the data. The retention is for abuse prevention.
196.
by usef- 4mo ago
You think companies would be ok with terms of service that allow potentially distributing their data and internal knowledge? It's an interesting question, though they tend to be more conservative than consumers
197.
by usef- 4mo ago
If that were dominantly true nothing would function at all. You trust and rely on thousands of people and services every day. As others have said, if you're this skeptical I don't see why you would have been using them before this
198.
by usef- 4mo ago
Sure, some trust is required that they aren't breaking their own terms of service (which legally enforces that they won't train on your data), but the same is true of every company/service you deal with (AWS, Google, your CRM
199.
by usef- 4mo ago
Maybe, but to do so they'd need to offer new terms of service and we'd have to accept. I believe they'd lose a lot of their core business market if they did so.
200.
by usef- 4mo ago
It's only for this model, not the one you're already using. And they're not training on the data. It's supposedly to detect abuse etc (such as someone retrying repeatedly with different variations to get around their pro
201.
by usef- 4mo ago
The phrase "security through obscurity" isn't an argument against all information restriction. It doesn't imply we should, for example, publish step-by-step instructions for making widespread death easier.
202.
by usef- 4mo ago
Compare the cost/ease of attacker vs defender if one person is given a virus to unleash anywhere in the world and another person is given a vaccine to distribute to the whole world. Or compare building a large bridge to someone disabli
203.
by usef- 4mo ago
The companies using distillation are still training on all that data too, aren't they?
204.
by usef- 4mo ago
What ladder did Anthropic use?
205.
by usef- 4mo ago
If I asked you over voice what Mike said he wanted for his birthday, it's usually less confusing if you just answer the question ("he wants snow boots") rather than started reading out words from another person ("I love
206.
by usef- 4mo ago
My mistake: it was Nasdaq that is being reduced to days, not S&P. Thanks.
207.
by usef- 4mo ago
S&P500 is not a total market index. It tracks a specific kind of large firm, with certain filters. Fast tracking means that the market likely wont have enough time to find the settled price (especially with the knowledge that passive fu
208.
by usef- 4mo ago
Oh yes, I know open source games existed, but did any raise enough to pay wages of a team? ps. I personally played Threes first, which 2048 copied without the style and made free. I admit I was slightly disappointed for the original devs th
209.
by usef- 4mo ago
Have there been commercially successful free software games? If the code was Free Software, how would you sell in-game purchases? Note that if it were just 3d models for your character, others could sell them too, likely cheaper without the
210.
by usef- 4mo ago
That sounds like a bug. They're very common for open model releases on the first day. If I wasn't on mobile I'd try it on Google's own app.

More ›