Grok-2 Beta Release
x.ai
x.ai
What is the company’s ethical position though? It officially stemmed from Mr Musk’s objection that OpenAI was not open-source, but it too is not open-source. It followed Mr Musk’s letter to stop all AI development on frontier models, but it is a frontier model. It followed complaints that OpenAI trained on tweets, but it also trained on tweets.
Companies like Meta, Mistral, or DeepSeek, address those complaints better, and all now play in the big league.
He is primarially known specifically as someone who is incredibly impulsive, is unable to differentiate fact from fiction and not actually interested in chasing any kind of objective truth in so much as that is possible.
But multiple times per week now for a long time you can see him sharing and commenting on things that are provably wrong and I don’t mean in some kind of “it’s just a different opinion” kind of way.
There is never any kind of introspection, never any kind of “oh I was wrong” just proceeds to roll immediately into the next round of bullshit.
So, no… people don’t have any kind of assumption that he has “strong ethics”. Maybe you meant strong convictions? Because that he certainly does have.
I'm sure you could make a case that these descriptors apply to him, not a particularly strong case, but ... You think he's primarily known for these things?
You will find yourself called unethical around different groups.
Understanding what group you are in is important to keep in mind when judging others.
No one is going to get investigated or have their careers ruined because a vegan called them a murder, obviously.
Unreal what sort of knots someone will tie themselves into to excuse this type of behavior.
Your worldview seems limited to tweets and social media in first world.
If you say something false in a context where it is likely to actually harm someone, or with the goal of actually harming someone, you're an asshole.
Your level of assholeness rises in tandem with the expected harm of your falsehoods.
His whole history tells otherwise.
- the pedo guy moment
- the hyperloop smoke and mirrors which are actually just his campaign against public transport
- the several union busting episodes
- the whole “Tesla is going to save the world” thing
- the multiple harassment cases
- the multiple instances of overwork, discrimination, and general lack of any consideration for his staff
- all the severance payments he failed to make after having fired a whole bunch of people
- the stupid ultimatum before one of the firing episodes, plus the utterly stupid “show your work” thing that came just after
- the multiple times he stiffed his creditors (either landlords, lawyers, contractors in general) or tried to do it
- the way he tried to force open Tesla factories in the middle of the COVID pandemic
- the multiple instances of pushing Russian propaganda verbatim (concerning.)
- most of the Neuralink saga
- the FSD vapourware that has been coming next year for a decade
- the way he publicly disparaged people who were killed by their Tesla using telemetric data that are supposed to be confidential
- the Media Matters lawsuit
Well, I could go on. He’d need to work quite hard to reverse his public image of massive arsehole at this point.
> the way he publicly disparaged people who were killed by their Tesla using telemetric data that are supposed to be confidential
In the cases I've seen, Tesla pulled data showing that the people "killed by their Tesla" were either not paying attention at all (contrary to Tesla's explicit warnings), or were driving without the automated features enabled after all despite initial media claims to the contrary. Is this what you consider "disparagement" or do you have more egregious examples?
That said, yes. What you said is evidence that he is mean-spirited and does not follow the rules he set himself. Of course, having an ethical behaviour sometimes means making hard choices. It is not about doing things that are understandable in context, it’s about doing the right thing, even if it is at a cost to you in the short term. He does not have any history of doing so.
These people are dead. Spitting on their graves because he is annoyed by their family is absolutely unethical. Particularly since their main failure was to believe the smoke and mirrors about FSD, which is itself another ethical clusterfuck.
If he had a beef, he could have sued for defamation, where he could have shown his data in an ethical and confidential manner. He knows the deal, he’s been in more than his fair share of defamation lawsuits, on either side.
But that's just me. I doubt further debate on this would be productive.
Oh he likes to bully people with lawsuits. Amber Heard comes to mind (he bullied the studio behind Aquaman).
He doesn't care if his defense of free-speech causes him revenue losses on X.
Having moral principles is completely orthogonal to being ethical. Ayn Rand had lots of moral principles and she was still a reckless sociopath. One of his moral principles is that greed is good, and his actions certainly are consistent with this one.
He did lose a lot of money on Twitter, but you can hardly call that him following his moral principles, considering how things actually happened.
> He doesn't care if his defense of free-speech causes him revenue losses on X.
Whose free speech is he defending? There is no evidence that he champions free speech, merely that he supports however agrees with him and edgelords. He is more than happy to harass, intimidate, bully, and be a general nuisance to those whose opinions he finds objectionable.
sounds like mr Musk is a conservative.
This is a little confusing in the US and other Anglo countries because traditionally we have been fairly liberal so sometimes people confuse liberalism and conservatism.
Without those, your comment is difficult to make any sense of, since the way you're using those words seems to differ from any sort of standardized definition.
Don’t get this confused with conservative and progressive politicians, though, who are generally ignorant of the actual traditions and philosophies behind their respective movements, and are essentially just cutouts for competing media and financial corporate interests. The few holdouts on both sides have been successfully sidelined (Bernie Sanders, Ron Paul), and it looks like the military-industrial complex will have their war with Iran no matter what the results of the next few elections are.
(Sorry, I’ll go have my coffee now and see if I get a little less doomer.)
Identity politics has always been a conservative project.
You know which ones I mean.
As an example: The "conservative" judge who threw out 40 years of precedent on a technicality to prevent the American public from learning whether their former and potentially future president sold, gave away, or otherwise exposed national security secrets after he undoubtedly stole those documents.
There's a fundamental asymmetry in "the movement" on the left - which essentially rounds out to whatever annoying undergrad student showed up in your Twitter feed today - and the actual elected, governing leaders of the right, doing things like throwing out very strong criminal cases on matters of deep public importance.
In other words, identity politics to a T.
An excellent example would be Murphy's Law, and by extension many of the similar, often eponymous, laws.
See:
- List of Eponymous Laws: <https://en.wikipedia.org/wiki/List_of_eponymous_laws>
- Murphy's Law and other reasons why things go wrong! by Arthur Bloch: <https://archive.org/details/murphyslawotherr0000arth>
- Compilation of Murphy's (and similar) laws: <https://www.cs.cmu.edu/~fgandon/miscellaneous/murphy/>
Some of those are humourous, some are in fact quite serious though have a comedic element particularly out of context. Most speak to at least a colloquial truth.
What Whilhoit did was manage to buttonhole a hypocrisy of modern conservativism, perhaps over the past few decades, perhaps a century or so (Anatole France, "The law, in its majestic equality, forbids the rich as well as the poor to sleep under bridges, to beg in the streets, and to steal bread", further evidentiarially supported by SCOTUS in Grants Pass), perhaps by millennia (see the opening paragraphs of A.H.M. Jones, Augustus, describing the political situation in the late Roman Republic, quoted here: <https://news.ycombinator.com/item?id=22208105>, and at greater length: <https://web.archive.org/web/20230607042525/https://old.reddi...>). It's not so much a proved hypothesis as a phrasing which fits the understanding of many and expresses it concisely and memorably.
And while I will grant you that liberalism (not to be confused with leftism), is different then conservatism, both (classical) liberalism and conservatism strongly require equal treatment (procedural symmetry).
Modern american conservatism very well fits the quote from grandparent.
Also, surely you would know what "repressive tolerance" is, since you're quoting it ? You would also know that the author you cite, whose name you misspelled, was critiquing the concept ?
And no, I have not seen a primary source where in context Trump said that he would "" "lock up" all the "sick, evil" democrats after he is reelected"". Do you have such a primary source? Years ago I was told that Trump said the white supremacists were "very fine people", so I looked at the transcript and he literally said the opposite.
I don't think he ever endorsed white supremacists, but I think I remember (wouldn't bet on it) an instance where a journalist asked him why some keep showing up at his rallies and why he does nothing about it. He then answered that he doesn't really know who they are, that he didn't know, etc. basically eluding the question. i.e. they're welcome but won't proclaim his support for them, one of the many dog whistles Republicans use nowadays.
This definition is plainly wrong on so many levels that it's practically impossible to engage with. But I'll make that mistake and engage on two points.
First, it implies that conservative position has somehow consistent features across time and space. There is difference between conservative in Germany, USA and China. Not to mention conservative in early, mid and late 20th century.
Second, ignoring legal norms is neither stated, nor implicated position of conservative political movements. At very worst, we can accuse them of maintaining laws with discriminatory intents. But not of flaunting those same laws.
The Republican candidate for president in the USA is a convicted felon
I say this as someone living in Pennsylvania, drowning in the hypocrisy and escalating hate this group of people has been spewing for the last ~8 years.
Therefore I can completely understand why someone might focus on that as the most relevant definition on 'conservatism' today in the USA.
You don't consider this a problem, that the word "conservatism" when discussed with an unknown recipient online (very possibly non-american) is constrained to the context of the past decade(s) in the United States?
Words have meaning, so if you're going to have a meaningful discussion about a word like "conservatism" or any type of -ism for that matter, I would think it benefits anyone engaging in that discussion to be aware of the different wings present in that word, whether that be across history or across present day geography.
https://slate.com/business/2022/06/wilhoits-law-conservative...
https://www.nationalreview.com/2018/07/roger-scruton-meaning...
You do not need to agree with him or his definition of Conservatism and there are other definitions of the term which are also applicable but none of those definitions have any resemblance to what you posed.
Musk is a centrist, not a conservative. He used to stand to the left of centre but has been moved to the right of centre by virtue of the left moving further left, thereby moving the centre to the left as well while Musk staid put.
Society needs conservatives just like it needs centrists and progressives and whatever other names you want to give to these philosophies and/or ideologies. A world made out of only progressives never gets anywhere since they will never find out what works well and what does not since their aim is to shape the future by means of societal change. A world made out of only progressives will eventually grow stale when the rate of change in the environment outpaces societies' capacity for change. Progressives can make good innovators but tend to be less able at keeping things running. Conservatives can be good at keeping things running but tend to be less inclined to innovation. These are broad brush strokes but the essence is sound, society tends to work best when there is a balance between conservatives and progressives.
You might notice some parallels: architects tend to be lousy builders, builders tend to be uninspired architects. Developers tend to be less gifted at UI design, UI designers tend to be sloppy developers. Copy editors tend to be unremarkable writers, writers tend to be less effective copy editors.
https://en.wikipedia.org/wiki/Francis_M._Wilhoit#:~:text=Con....
> However, it was actually a 2018 blog response by 59-year-old Ohio composer Frank Wilhoit, years after Francis Wilhoit's death."
- xAI opens sources models with a 6 month lag, look at Grok 1
- No one else stopped development, so why should he?
- He owns Twitter, why wouldn't it be okay for him to train on Tweets?
They tried to add a pre-checked mark to the settings, but at least in Europe, where we actually have consumer protection, that won't fly.
That's what happened once, rather than a policy that we can expect to be applied. (Unless I missed some announcement?) Based on "we'll publish the algorithm" which ended up being a one-off partial snapshot, never updated afterwards, I wouldn't hold my breath for the models.
> He owns Twitter, why wouldn't it be okay for him to train on Tweets?
There's a whole thing about having clear opt-in agreement about how your data will be used for EU citizens. Twitter didn't comply here with their hidden opt-out strategy.
Because he doesn’t own the tweets. Can you imagine if posting a photo you took to Twitter meant it’s not your photo anymore? Totally ridiculous.
You retain your rights to any Content you submit, post or display on or through the Services. What’s yours is yours — you own your Content (and your incorporated audio, photos and videos are considered part of the Content).
By submitting, posting or displaying Content on or through the Services, you grant us a worldwide, non-exclusive, royalty-free license (with the right to sublicense) to use, copy, reproduce, process, adapt, modify, publish, transmit, display and distribute such Content in any and all media or distribution methods (now known or later developed). This license authorizes us to make your Content available to the rest of the world and to let others do the same. You agree that this license includes the right for Twitter to provide, promote, and improve the Services and to make Content submitted to or through the Services available to other companies, organizations or individuals for the syndication, broadcast, distribution, promotion or publication of such Content on other media and services, subject to our terms and conditions for such Content use. Such additional uses by Twitter, or other companies, organizations or individuals, may be made with no compensation paid to you with respect to the Content that you submit, post, transmit or otherwise make available through the Services.
In fact, they didn't include any purpose for their own use of the data and following GDPR thus cannot use the data at all. They did include purposes for other companies (syndication, broadcast, etc) which also doesn't include training of AI.
“you grant us a worldwide, non-exclusive, royalty-free license (with the right to sublicense) to use, copy, reproduce, process, adapt, modify, publish, transmit, display and distribute such Content in any and all media or distribution methods (now known or later developed).“
Not sure how anyone could defend that an AI model is not covered by this idea - such a model is easily covered by “distribution methods”.
Adding a new purpose requires additional consent at least in the EU.
If it is within his right to use this data for training purposes, then that's it.
And he is, btw.
And those terms were in place since way before he took over Twitter, btw, btw.
I thought it was a moral imperative or some such thing to do AI right because it could "destroy humanity"?
Or was that just Musk and the rest of the special people in SV's way of aggrandizing themselves while trying to do something most of them have either no experience in or fail miserably at, which is raise an intelligence to be a responsible actor?
In a recent interview on Lex Fridman, he envisioned a future where humans augmented AI, through a brain to computer device like Neuralink would be able to keep up with pure AI.
Now one can immediately notice a hole in this reasoning: namely what guarantees that the AI that is use to augment humans is going to be benevolent and won't go rogue?
You don't need an "aligned" AI to tell you the distance between the Earth and the Moon. You need an "aligned" AI to tell you not to rape people even if you can get away with it, that's because the idea that rape is bad is not an objective truth based on the laws of nature, it is a "social consensus".
Truth isn't one fundamental thing, truth is what works.
There are places in the world today that the social consensus is it's fine for a man to rape his wife. Social systems in these places don't work very well, and one can make a logical and well reasoned argument linking the social acceptance of rape to a myriad of other dysfunctions.
If this isn't your point, what is? Just that you don't trust Elon?
Do I think rape is good? Absolutely not, because I follow the current social consensus on that one, not a "truth" that is muddy at best. And I also want AIs to do the same.
So I guess you would have been fine with being a concentration camp guard, rounding up Jews and putting them in the oven because social consensus said it was the right thing to do?
Maybe you see the logic as shaky because you lack knowledge in logic and ethics...?
Ethics gives us more questions than answers. The trolley problem doesn't have a true answer for instance.
The human rights are a social consensus, it is even made explicit by being a signed declaration. It felt good to the people who wrote them, it also feels good to me, because I was born and live in a society that has these values. It is only truth because by social consensus, we decided it is. In logic that would be an axiom and aligning an AI would mean implementing social consensus as axioms.
There are some fundamental reasoning that can justify human rights. One can use game theory, or the idea that human rights promote free thinking and free thinking is what brings the most value out of people now that machines do better than slaves for menial labor. But these are, I think, not enough.
Absolutely NOT social consensus as axioms, as that will result in stagnation and tyranny.
Instead we must progress gradually one axiom at a time through reasoning and experimentation.
Truth is not what we decide it is, truth is what works. The universe decides what is true, not people.
Re your comment: "but these are, I think, not enough". I both agree and disagree depending on what you mean. Fundamentally this approach is enough, but practically we haven't developed our understanding enough to map out absolute truth. It's probably something we can only approach but never reach.
But in theory the right AI system could allow us to approach the faster
Not only that - Grok is/was trained on ChatGPT output, which I suppose Musk felt was turnabout. When asked about its identity, the first Grok would respond like ChatGPT (https://news.ycombinator.com/item?id=38584922)
Only Meta’s Llama can be used for work. The rest are just toys for personal use noobs who don’t read the fine print.
Since these "safety" features tend to also degrade the model, that's likely also helping them catch up in the benchmarks.
This further feeds into my concern that the more advanced AI models we get, random enthusiasts at that site may no longer be able to rank them well, and tuning for Chatbot Arena might be a thing. One that is also exploited by GPT-4o. GPT-4o absolutely does not rank wildly ahead of Claude 3.5 Sonnet in a wide variety of benchmarks, yet it does in Chatbot Arena... People actually using Claude 3.5 Sonnet are also quite satisfied with its performance, often ranking it more helpful than GPT-4o when solving engineering problems, but at the expense of tighter usage limits.
Chatbot Arena was great when they were still fairly stupid, but these days, remember that everyday people are put against the task of ranking premium LLM's even solving some logic puzzles, trick questions and with a deep general knowledge far beyond that of singular humans. They can strike against traditional weaknesses like math, but then all of them suffer. So it's not an easy task at all and I'm not sure the site is very reliable anymore other than for smaller models.
One thing is sure - that current commonly used benchmarks are mostly polluted and worthless. So you have to go to niche ones.
For example the one I check for coding is Aider LLM leaderboard [1].
We maintain Kagi LLM Benchmarking Project [2] optimized for the use case of using LLMs in search.
You can review for yourself and decide if it was justified (you can compare based on W/L/T responses and matchups). Generally, Claude still has more refusals (easy wins for the model that actually answers the request), often has worse formatting (arguable if this is better, but people like it more), and is less verbose (personally, I'd prefer the right answer with less words, but ChatArena users generally disagree).
If you look at the questions (and Chat Arena and Wildchat analyses), most people aren't using LLMs for math, reasoning, or even coding - if anything the arena usage is probably overly skewed to reasoning/trick questions due to the subset of people poking at the models.
Of course, different people value different things. I've almost exclusively been using 3.5 Sonnet since it came out because it's been the best code assistant and Artifacts are great, only falling back to GPT-4o for occasional Code Interpreter work (for tricky problems, Mistral's Codestral actually seems to be a good fallback, often being able to debug issues that neither of those models can, despite being a tiny model in comparison).
Two more things concerning Chatbot Arena:
- The prompts people use on have an incredible sample bias towards certain tasks and styles, and as such are unrepresentative of "overall performance" which is what people expect from a leaderboard.
- It is incredibly easy to game by a company, their employees or their fanboys if they would like to. No idea if anyone has done so, but it's trivial.
Just to give one example of the bias; advances in non-English performance don't even register on the leaderboard because almost everyone rating completions there is doing so in English. You could have a model that's a 100 in English and a 0 on every other language, and it would do better on the leaderboard than a model that's a 98 in every human language in the world.
Have you tried putting in "photorealistic" instead of "realistic", assuming that's what you mean? I'm curious if that would get better results.
When they make the mini model available for download and quantizable. That's when I may be interested. But given the minimal improvement in the past several months, I'm inclined to believe that we have reached the plateau.
This is Musk after all, so I wouldn't be surprised if it strayed far from the norm.
Do you guys have any idea how sinister "safety" sounds in this context?
But I too wouldn't use this. X is playing fast and loose with ... everything, so having a business rely on their product seems risky.
With LLMs (and AGI) it's really that simple: the company with the best model wins regardless of all else.
It's going to be a combination of price, performance, quality, reliability, availability etc.
And since the prompts need to be optimised for each model there is a degree of vendor lock-in.
I'm talking about the next huge step forward that only 1 company will achieve, because it simply has the most GPUs (in limited supply) + energy source first and keeps that advantage.
At some point this becomes a run-away self amplifying differentiator and it will make that company win regardless of all else.
My money is on xAI in 2025.
PS: the only reason prompts need to be optimized for each model is a symptom of models simply not being that good yet. This need will vanish in the near future as you get way better models. A recent hint of what I mean: mid-journey needed very elaborate prompt (and even loras) to get what you want. In flux that prompt can be much shorter (without loras) and it still gets closer to what you want. Same will happen with LLMs. Another example: with ChatGPT 4 you need to literally beg a model to only return what you ask for (for example JSON) or put it in a certain mode (JSON mode), in Claude Sonnet 3.5 it will simply just listen to what you ask for. So again: that's not "because every model needs model-specific fine-tuning" that's because previous models where simply not as good.
Sometimes having a fast enough model at a low enough price makes you the obvious choice e.g. I know Claude is better than gpt-4o-mini but I use the latter for a lot more data processing because it's significantly cheaper and faster and the gains I'd get out of Claude seem somewhat marginal for my use case
Best at product / market fit. And that space is very very wide. Does the GenAI serve as a feature in a larger product (like realtime “reasoning” on X or in Apple’s case in iOS)? Is it a standalone product that general public or enterprises use? Does it play in a niche area? Etc.
https://www.theverge.com/2024/8/12/24219121/donald-trump-elo...
If that's true it's not exactly the sort of behaviour you want from an API you're depending on.
Evidence for DDOS:
- Elon said so
- the event in question very clearly had huge technical issues
Evidence Against DDOS:
- Elon said so
- People who worked at Twitter said it was bullshit
- every other spaces event that was run at the same time was unaffected.
- no other part of the website was impacted in any way whatsoever.
Aren’t these last two an argument FOR a ddos attack? It seems reasonable to assume we’re there a ddos attack at that time it would be against the Elon/Trump stream explicitly.
So no I think it was just a straight up technical failure on their end.
No, we have no idea from The Verge article whether the sources are even qualified to make such statements or if the statements are even true. In fact on the basis of the 99 percent speculative quote we can disregard the source quotes altogether. I'll say this, I work on far less significant software than X and we get DDOSed all the time.
> every other spaces event that was run at the same time was unaffected.
That's not true, I wasn't even able to load my feed during the initial part of the stream.
You baselessly accuse journalists of straight up making things up and then go on to give some anecdotal evidence that conveniently nobody can disprove.
The Register has found no evidence of a denial of service attack directed at X. Check Point Software's live cyber threat map does not record unusual levels of activity at the time of writing. NetScout's real-time DDoS map recorded only small attacks on the US.
If a DDoS was indeed the reason for the delayed start of the event, it appears not to have impacted the rest of X's operations – there were plenty of posts commenting on the problems with the Space occupied by the interview. And Musk was tweeting from the very network said to be under attack.
They also threw shade on the numbers:
The interview commenced some 40 minutes after its advertised time. Live audience statistics reported 1.1 to 1.3 million attendees during the portions of the event The Register observed – although during the stream Trump claimed that the event had an audience of 60 million or more, exceeding targets of 25 million.And unfortunately both of these men are known for bullshitting more than anything else and have been now for a long time.
Was Trump a fool to count the people that took one look and changed channels, or a knowledgable and deliberate deceiver?
The Verge has no political bias, has a good reputation and thus deserve the benefit of the doubt.
If they were actually neutral, they'd phrase it more like: with technical difficulties.
We know nothing about the sources, and writers are not above making stuff up. I could just as easily spin it on them: there's a 99 percent chance they made up the sources.
From a black box perspective, LLMs are pretty simple, you put text or images in, (possibly structured) text comes out, maybe with some tool invocations.
If you use a good library for this, like Python's litellm for example, all it takes is changing one string in your code or config, as the library exposes most APIs of most providers under a simple, uniform interface.
You might need to modify your prompt and run some evals on whatever task your app is solving, but even large companies regularly deprecate old models and introduce vastly better ones, so you should have a pipeline for that anyway.
These models have very little "stickiness" or lock-in. If your app is a Twitter client and is built around the Twitter API, turning it into a Mastodon client built around the Mastodon API would take a lot of work. If your app uses Grok and is designed properly, switching over to a different model is so simple that it might be worth doing for half an hour during an outage.
Trying it isn't exactly locking you into anything
I'm surprised they managed to catch up. I guess there really is no moat.
My guess is that they're using one of the third party AI training outfits for this and that they are paying through the nose.
This looks exactly like a training task I got to see on one of those platforms.
I’m not hugely optimistic, though.
the more valuable part is the dataset which probably requires a lot of people to hand-filter
and even more impossible is acquiring the training rig, which your average person can't even afford
yet to see if this new model is able to do any better on that regard
And also remember to vote in the next EU elections pro development, not against development, parties.
Is anyone with X premium able to confirm the vibe check -- Is the model actually good or another case of training on benchmarks?
So results are mixed, but the real takeaway is this is a competitive model that is good enough to be worth using today. It puts xAI significantly above their previous position, and up near the top of the field with OpenAI, Anthropic, Google, and Meta. And their new H100 cluster should allow them to keep up with the next wave of releases, whenever that starts.
Right now I'm not really a believer in Grok and I doubt it will be worth using.
More importantly, if the model is not a huge leap at this point I just don't care if it is as good as the very limited models we already have because I am not impressed by any of these anymore.
Anything less than a 3.5 to 4 jump from here is just not going to vibe for me.
That said I’m still cheering for mistral and meta with their more open stance
https://noyb.eu/en/twitters-ai-plans-hit-9-more-gdpr-complai...
"He used one of my eggs to irreversibly make a cake"
It's true, but it would be kind of amazing if it weren't
From this knowledge, you can ask the AI "Who has driven to Mexico" and it might know that Ben Adams did, and reply with that.
HOWEVER it's also baked into the model and can't be surgically removed after a complaint. That's the irreversibility part. You can't undo isolated training. You need to provide it a new data set and train it all over again. They won't do that because it's too costly.
The problem with the above example is of course that it can also contain sensitive or private user details.
I've easily extracted the complete song lyrics to the letter from GPT-4 even if OpenAI try to put up guardrails against it due to the copyright issues. AI is really still in the wild west phase...
"He used his eyes to irreversibly read this post"
Additional Twitter data is in my eyes mostly low quality content, that's nothing I would want in a AI model.
That may be considered a feature.
ChatGPT seems reasonably concise, Gemini's answers tend to be verbose (without adding meaningful content).
In particular, requesting an analysis of the problem first before jumping to conclusions can be more effective than just asking for the final answer directly.
However, this analysis phase, or similar one, could just be done hidden in the background, but I don't think any are doing that yet. From the user point of view that would be just waiting, and from API point of view those tokens would also cost. Might just as well entertain the user with the text it processes in the meanwhile.
That doesn't seem to be the case any more and there has been speculation this is down to the star method being used for training newer models. I say speculation because I don't believe people have come out and said they are using star for training. OpenAI referred to Q* somewhere but they wouldn't be drawn on whether that * is this "star" and although google were involved in publishing the star paper they haven't said gemini uses it (I don't think).
How does it matter even if the quality is high or low? The point is user data was used without consent.
(But regardless, many people raised an issue of OpenAI training from sources they shouldn't be allowed to access, so they're definitely a problem as well)
They provide a product in the EU, therefore they must either follow EU law or exit the EU market. Just like an EU company that provides a product in the US has to follow US law.
The line of 'following the law of another country' is grey area on the internet, given that it goes both ways:
EU online companies providing services to US users fail to provide the free speech guarantees that the US laws afford their citizens. That's because all EU countries have more strict laws limiting free speech. Should the EU companies break their own countries' law to satisfy the US audience?
Could you sharpen up this claim? Like suppose I run a microblogging site but I delete libellous posts and incitements to violence in accordance with my local European law. Am I violating a US law by allowing Americans to use the site?
Exactly what is "free speech guarantees" in the context of a private business?
So it seems there are states where a europeans social medium should abide by rules that would most likely contradict european laws, right?
i didnt test any crazy promps on grok but the image generation won't let you do things like naked people for example
PS: I have X Premium and I'm quite liking Grok 2.
And the primary way to evaluate the worth of a social network is user engagement metrics.
I had to resort to configuring "muted words" to actually fix the problem.
Its very weird to me that such basic things seem so bugged on a platform that popular, but other much much more complex things (mass video streaming, Grok itself) work totally fine?!
Lamma 3.1 is competitive.
Not perfect but better than open EU citizens to global data mining.
Is that based on any facts or just the typical EU myths like the cucumber regulation?
Countries with the highest average standard of living: Check.
Countris consistently socring among, or as, highest in citizen happiness indices: Check.
Stable Economy: Check.
Successfully implemented measures to combat a global pandemic: Check.
Slowly but surely fading out dependency on russian fossile fuels: Check.
Best privacy protection laws in the world: Check.
Shall I go on?
I asked about the brussels bubble. Only the privacy laws are relevant. I can't say that it made a dent to our overall privacy, since our tech is US based and we have no idea where our data really are.
These countries are doing great because they are part of the EU. If anyone disagrees, well, it's not like we lack experimental data:
https://www.london.gov.uk/new-report-reveals-uk-economy-almo....
https://www.gisreportsonline.com/r/brexit/
https://www.politico.eu/article/political-gridlock-northern-...
The safest countries in the world are in East Asia; there's no country in Western Europe where you can leave your wallet on the table in a major city and not worry about it being stolen.
>Stable Economy: Check.
It's stable in the sense that for most of western Europe (France, Italy, Spain, Portugal, Greece) GDP per capita now is no higher than it was 10 years ago. One of the few places in the world where people's material standard of living is no longer improving every year.
Denmark is a member state since 1973, when it was still called the European Economic Community
https://en.wikipedia.org/wiki/Denmark#Constitutional_monarch...
Denmark, together with Greenland but not the Faroe Islands, became a member of what is now the European Union, but negotiated certain opt-outs, such as retaining its own currency, the krone.
And Iceland is part of the European Economic Area (EEA)
https://en.wikipedia.org/wiki/Iceland
Iceland joined the European Economic Area in 1994, after which the economy was greatly diversified and liberalised.
And no, this list showcases mostly countries that are in fact EU members: https://en.wikipedia.org/wiki/Member_state_of_the_European_U...
Cookie popups on every website, its completely idiotic. Ofcourse I end up wasting hours of my life clicking on them randomly. There should just be an OPT IN for the cookie popup, or OPT OUT. Either on a EU portal or something.
Infact I find it fascinating noone is coming up with a better implementation of privacy concerns in the EU. This whole US not releaseing software incase they get sued, and thinking its fine for the whole EU population to go around clicking popups every 2 seconds is a total failure that needs to be fixed.
Again the idea is good, but the implementation needs to be fixed.
If you really have to experience Musk's chat bot on day one just use a VPN?
That's not because of the EU, that's because the website owner try to annoy you to blame the EU.
If they stop tracking no pop up is needed.
>Its more important for me to have access to cool tech
It's hardly possible to only give your own data without exposing third parties who may be not to keen of sharing their data.
FB already showed that. I bet WhatsApp already has my phone number without my consent because some relative of mine uploaded his contacts.
You'll have to say goodbye to BBC iPlayer though.
Im just saying the regulatory landscape seems to be changing in many ways and there are many reasons why people would consider emigrating, not just the fact they have to wait months for tech access in a rapidly changing technical landscape.
also be careful, you might be arrested for these comments in a few years time (lol)