Three major LLM releases in 24 hours
simonwillison.net
simonwillison.net
Still 281GB is huge. That's at the higher end of what we see from other open weight models, and it's not going to fit on anybody's homelab franken-GPU rig. Assuming that 281GB is fp16, it should quantize down to roughly 70GB at 4bits. Still too big for any consumer grade GPU, but accessible on a workstation with enough system ram. Mixtral 8x7B runs surprisingly fast, even on CPUs. Hopefully this 8x22B model will perform similarly.
EDIT: Available here in GGUF format: https://huggingface.co/MaziyarPanahi/Mixtral-8x22B-v0.1-GGUF
The 2-bit quantization comes to 52GB, so worse than my napkin math suggested. Looking forward to giving it a try on my desktop though.
Of course you might not care, if you believe that the EU has no way of forcing you to pay a fine and if you are certain that you are never going to do any business in the EU. But in such case you might as well provide access to your models for EU IPs too. It makes no difference.
The model access makes no difference.
"Personal data" and "data processing" are (deliberately) drawn very broadly:
"""Personal data — Personal data is any information that relates to an individual who can be directly or indirectly identified. Names and email addresses are obviously personal data. Location information, ethnicity, gender, biometric data, religious beliefs, web cookies, and political opinions can also be personal data. Pseudonymous data can also fall under the definition if it’s relatively easy to ID someone from it.
Data processing — Any action performed on data, whether automated or manual. The examples cited in the text include collecting, recording, organizing, structuring, storing, using, erasing… so basically anything."""
And, unlike the arguments about copyright in big AI models trained on the internet (is it 'fair use'? Don't ask me, IANAL!), the requirement for explicit and informed consent is something a general crawl will very clearly fail:
"""Purpose limitation — You must process data for the legitimate purposes specified explicitly to the data subject when you collected it."""
Furthermore, we don't know enough about how the models store knowledge/beliefs to be able to make any claim about accuracy:
"""Accuracy — You must keep personal data accurate and up to date."""
And as for confidentiality… for downloadable models, that's "by obscurity" only, due to the exact same research needed to resolve the previous point about accuracy, and even for secret models like GPT-4, nobody's really sure how to actually guarantee it won't leak info with the right prompt, and there's even some suggestion that this is actually impossible with current approaches because nothing is really deleted by RLHF:
"""Integrity and confidentiality — Processing must be done in such a way as to ensure appropriate security, integrity, and confidentiality (e.g. by using encryption)."""
In the case of #1, they probably do not have an agreement with the "data controller" in the case of scraping, which means #2 is a violation of GDPR.
IANAL.
As a European I to try see it from both sides, consumer protections are generally a good thing, but it right now being restricted by EU vagueness sucks ass because I just want to play with the cool new toys.
They could have easily made law requiring sites accept DNT header but they didn't likely because of lobbying.
DNT is not relevant as GDPR is not directly a regulation against tracking, and it certainly isn't because of lobbying.
> If you want to comply it is easy.
That's the entire antithesis of modern law as opposed to monarchy. Law should be codified in as clear rules as possible.
If by design you can do whatever pleases you, then yes you have a lot of innovation. But sometimes it leads to normalisation of troubles (e.g. data leaks in the US), and incredulity of the general public ("how did we even get there")?
There are good reasons to ponder ethics in the original balance too, it hasn't got to be completely paralysing either. But this comes with a cost (e.g. typically, for any data-sensitive work these days complying with GDPR, a significant part of the design & implementation time is "are we compliant").
What is OpenAI GPT4 and Google bard/Gemini EU version not doing so they work in EU but the latest Google AI is doing so google is incapable of putting it in EU ?
Maybe latest one is more invasive with ypur personal account? Scanning your personal data without consent ?
Because seems to be a super simple business, user sends you a prompt, you run the prompt and send the result back, you keep nothing unless necessary for the service to work and give the user the ability to purge their history/data .
I'm glad that Microsoft / OpenAI is confident enough to put out a global consumer facing gen AI offering. Maybe they received EU assurances, or maybe they have enough legal muscle to tank any possible hit, and eating Google's lunch is a prize worth the risk.
Google is in a completely different situation. Gemini is, so far, a niche product, making little to no money for Google. However, due to how the GDPR works, a mistake with Gemini could cost the entire company dearly, impacting the profits they make from any other products.
This is a more general principle. If you're a small startup that manufactures foos, and there's a risk that manufacturing Foos infringes on Apple's patents, well, that's just one of the risks you have to bear as a startup. If you get wiped out by a lawsuit, you're established as a limited-liability corporation for a reason, if the fine is larger than the worth of your company, you can just declare bankruptcy and go away. You can't decide to stop manufacturing Foos, as that's your entire business.
If you're the size of Google, getting into the Foo-manufacturing business is not worth it. If Apple sues you, that won't just impact the Foo-manufacturing side of your business, the fine can be significantly larger than any profits you ever hoped to make from the venture, not to mention the brand damage, strain on resources etc. If Foo manufacturing is ultimately going to be a side business for you, it's probably not worth it to be the hill you die on.
This is why it makes sense for startups to "move fast and break things™", and for big corporations to require countless legal reviews on the smallest of decisions.
I wouldn't be surprised if there's a long and arduous procedure to clear a new product for GDPR compliance, and the PM had a choice between launching now and excluding the EU or launching later for everybody. This is also what originally happened with Bard.
Lots of companies come unstuck because they fall into the trap of “let’s just collect everything and see what we can do with it”.
Or, I’ve got all this data I’ve collected legitimately. Who knew that you could sell it on to some data broken and make loads of money - let’s do that!
Or, I’ve collected all this data, I’m just going to keep it hanging around, oops I just put it on a public bucket and leaked it all. Hmm, I’m not even sure what data we had, have we just compromised a bunch of people? Who knows…
It's not supprising that many smaller companies are saying fuck this noise.
* It's complicated so it takes a while and you need lawyers and such to make it right
* Rules for training are probably hugely vague and undefined. Because you could ingest personal data and it cannot be deleted
* AFAIK it needs to be hosted in Europe (not directly GDPR related, but america has laws that allows them to spy on all traffic in the US, so this is somewhat the counter to that)
In the end from my experience just working at a company that needs to be compliant this usually means:
* All the services need to be hosed in EU including 3rd parties we send any data to
* There needs to be a way (email is enough) to delete user data (including from 3rd parties which need an endpoint so you can trigger it from your side)
* You need to inform the user about the data useage and allow them to opt out of the "usage" of this data for non-essential things (i.e marketing emails). This does not mean you cannot save this data if you also use it for other things, but you can not use it for the non-essential case.
* You could be in trouble if you save data "just because" and do not use it for anything essential or if it is not transparent to the user.
Not a lawyer. Just the things I notice in my day to day. In the end companies need data protection professionals to navigate these things. Which is probably another thing a startup does not worry about it early on.
If I'm to venture a guess, it's probably because data protections are stronger and they want to avoid potential issues should someone test GDPR (or whatever the applicable law is) by asking specific data be removed from the model
They are literally doing the opposite because they are asking you to commit to the terms and since everyone do you have actually consented to your data being used.
AI Act isn't solving anything that isn't already solved with existing regulation.
That so many people on HN seem to think this is a good idea is very puzzling.
What is the downvote coming from, isn't this just facts? If not for the regulation, why would EU be shunned?
I am using gpt-4-1106-preview quite a lot, but it is hard to optimize prompts when you cannot build a test-suite of questions and correct replies against which you can test and improve the instruction prompt. Even when using temperature=0, gpt-4-1106-preview outputs different answers for the same prompt.
[0] https://platform.openai.com/docs/api-reference/chat/create#c...
'temperature': 0,
'seed': 1,
The 'system_fingerprint' in the reply was the same in both of the json responses. So it seems that even when you get the same 'system_fingerprint' back, replies for the same prompt will not be the same.
I think this is because your approach isn't right. This tech isn't really unit-testable in the same sense. In fact, for many use cases, you may want non-deterministic results by design.
Instead, you probably need evaluations. The idea is that you're still building out "test" cases, but instead of expecting a specific result each time, you get a result that you can score through some means. Each test case produces a score, and you get a rollup score for the suite, and that's how you can track regressions over time.
For example, in our use case, we produce structured JSON that has to match a spec, but we also want to have the contents of that valid-to-spec JSON object be "useful". So there's a function that defines "usefulness" based on some criteria that I've put together since I'm a domain expert. This is something I can evolve over time, using real-world inputs that produce bad or unsatisfying outputs as new evaluations for the evaluation suite.
Fair warning though, it's not very easy to get started with, and there's not a whole lot of information about doing it well online.
Comparing the scores of two prompts will not give you a definitive answer which one is superior. But the prediction which one is superior would be better without the noise added by the randomness in the execution of the LLM.
Yes, but it can tell you which is likely to be superior, which is perhaps good enough?
Offline evals are only a part of the equation though, which is why online evaluations are perhaps even more important. Good observability and a way to systematically measure "what is a good response" on production data is what ultimately gets us closer to real truth.
I don't know why, but it is easy to imagine a scenario where non-determinism makes sense. For example the computer could say "I am supposed to do this floating point division. Let's see ... my very exact FP-Division unit is pretty hot at the moment. 82°C. Let's pass this calculation to the slightly less exact FP-Division unit then".
i think the kind of scenario you're describing does not occur in commercially sold hardware, because it poses problems not only for customers but also for verification engineers at the design house. also, in the specific case of floating-point division, it would violate the ieee-754 standard. it might happen in an experimental academic project
I like that both OpenAI and Anthropic default to the prepaid mode; I can safely experiment without worrying about selecting a large file by mistake (or worse, a runaway automated process).
Mixtral 8x22B is interesting because 8x7B was one of the best (among all others) for me few months ago (in particular, common knowledge, engineering and high-level math, multi-lingual skills like translation, grammatically nicer rewritings)
The reason to be able to migrate is for me personally a huge plus.
I never have found a need for them. i.e. the example in the article Just prompting like:
Write hello 3 different ways in spanish works fine for me
If this was 100% robust then it would also solve prompt injection, but sadly it isn't.