I really struggle to see the usecase of Grok Code Fast when you have Qwen 3 Coder right there providing much better outputs while still being fast and cheap.
On the plus side, GPT5 is very malleable, so you CAN prompt it away from that, whereas it's very hard to prompt Claude into producing hard code: even with a nearly file by file breakdown of a task, it'll occasionally run into an obstacle and just give up and make a mock or top implementation, basically diverge from the entire plan, then do its own version.
Absolutely, sometimes you want, or indeed need such complexity. Some work in settings where they would want it all of the time. IMHO, most people, most of the time don't really want it, and don't want to have to prompt it every time to avoid it. That's why I think it's still very useful to build up experience with the three frontier models, so you can choose according to the situation.
Haha, this does make me laugh a bit. They've not been doing antything illegal (they almost certainly have fwiw) because they've been the ones capturing regulations and massive influencing laws.
Furthermore they've violated plenty of consumer protection laws in countries where those are strong, so the entire premise doesn't hold up.
I meant property rights violations generally. The consumer agreed to something which they didn’t provide. If the consumer simply wasn’t served better than he could have, that may violate some non-objective law, but not the consumer’s property rights.
Their violations is the same kind of persecution that minority group face, such as with gay marriages. The laws are written specifically to target them.
Influencing the law can be quite different. Lobbying is legal, other kinds of peddling are not. If they exceeded their legal limits (or whatever, I’m not into it) and, in fact, took the role of an aggressor, that would be a problem and the government would have to sue them. Even in this case, destroying them should be the very last resort, not the norm.
Can't agree with that. Gemini doesn't lead just on price/performance - ironically it's the best "normie" model most of the time, despite it's lack of popularity with them until very recent.
It's bad at agentic stuff, especially coding. Incomparably so compared to Claude and now GPT-5. But if it's just about asking it random stuff, and especially going on for very long in the same conversation - which non-tech users have a tendency to do - Gemini wins. It's still the best at long context, noticing things said long ago.
Earlier this week I was doing some debugging. For debugging especially I like to run sonnet/gpt5/2.5-pro in parallel with the same prompt/convo. Gemini was the only one that, 4 or so messages in, pointed out something very relevant in the middle of the logs in the very first message. GPT and Sonnet both failed to notice, leading them to give wrong sample code. I would've wasted more time if I hadn't used Gemini.
It's also still the best at a good number of low-resource languages. It doesn't glaze too much (Sonnet, ChatGPT) without being overly stubborn (raw GPT-5 API). It's by far the best at OCR and image recognition, which a lot of average users use quite a bit.
Google's ridiculously bad at marketing and AI UX, but they'll get there. They're already much more than just a "bang for the buck" player.
FWIW I use all 3 above mentioned on a daily basis for a wide variety of tasks, often side-by-side in parallel to compare performance.
My pet theory without any strong foundation is because OpenAI and Anthropic have trained their models really hard to fit the sycophantic mold of:
===============================
Got it — *compliment on the info you've shared*, *informal summary of task*. *Another compliment*, but *downside of question*.
----------
(relevant emoji) Bla bla bla
1. Aspect 1
2. Aspect 2
----------
*Actual answer*
-----------
(checkmark emoji) *Reassuring you about its answer because:*
* Summary point 1
* Summary point 2
* Summary point 3
Would you like me to *verb* a ready-made *noun* that will *something that's helpful to you 40% of the time*?
===============================
I suspect this has emerged organically from the user given RLHF via thumb voting in the apps. People LIKE being treated this way so the model converges in that direction.
Same as social media converging to rage bait. The user base LIKES it subconsciously. Nobody at the companies explicitly added that to content recommendation model training. I know, for the latter, as I was there.
Gemini does the sycophantic thing too, so I'm not sure that holds water. I keep having to remind it to stop with the praise whenever my previous instruction slips out of context window.
Oh god I _hate_ this. Does anyone have any custom instructions to shut this thing off. The only thing that worked for me is to ask the model to be terse. But that causes the main answer part to be terse too, which sucks sometimes.
Not the case with GPT-5 I’d say. Sonnet 4 feels a lot like this, but the coding and agency of it is still quite solid and overall IMO the best coder. Gemini2.5 to me is most helpful as a research assistant. It’s quite good together with google search based grounding.
Gemini does this too, but also adds a youtube link to every answer.
Just on the video link alone Gemini is making money on the free tier by pointing the hapless user at an ad while the other LLMs make zilch off the free tier.
I've experienced the opposite. Gemini is actually the MOST sycophantic model.
Additionally, despite having "grounding with google search" it tends to default to old knowledge. I usually have to inform it that it's presently 2025. Even after searching and confirming, it'll respond with something along the lines of "in this hypothetical timeline" as if I just gaslit it.
Consider this conversation I just had with all Claude, Gemini, GPT-5.
<ask them to consider DDR6 vs M3 Ultra memory bandwidth>
-- follow up --
User: "Would this enable CPU inference or not? I'm trying to understand if something like a high-end Intel chip or a Ryzen with built in GPU units could theoretically leverage this memory bandwidth to perform CPU inference. Think carefully about how this might operate in reality."
<Intro for all 3 models below - no custom instructions>
GPT-5: "Short answer: more memory bandwidth absolutely helps CPU inference, but it does not magically make a central processing unit (CPU) “good at” large-model inference on its own."
Claude: "This is a fascinating question that gets to the heart of memory bandwidth limitations in AI inference. "
Gemini 2.5 Pro: "Of course. This is a fantastic and highly relevant question that gets to the heart of future PC architecture."
Not really. Any prefix before the content you want is basically "thinking time". The text itself doesn't even have to reflect it, it happens internally. Even if you don't go for the thinking model explicitly, that task summary and other details can actually improve the quality, not reduce it.
I recently started using Open WebUI, which lets you run your query on multiple models simultaneously. My anecdote: For non-coding tasks, Gemini 2.5 Pro beats Sonnet 4 handily. It's a lot more common to get wrong/hallucinated content from Sonnet 4 than Gemini.
Agreed. People talk up Claude but every time I try it I wind up coming back to Gemini fairly quickly. And it's good enough at coding to be acceptably close to Claude as well IMO.
Google also has a lot of very useful structured data from search that they’re surely going to figure out how to use at some point. Gemini is useless at finding hotels, but it says it’s using Google’s Hotel data, and I’m sure at some point it’ll get good at using it. Same with flights too. If a lot of LLM usage is going to be better search, then all the structured data Google have for search should surely be a useful advantage.
I don't trust people saying that "society is going down the abyss" and then using it to justify a crackdown on personal freedoms - ranting about "degeneracy" is how authoritarians destroy democracy time and again.
“Freedom” in the American context means something different than the how people use it today. It’s closer to “freedom to make the right choices.”
John Adams said: “Our Constitution was made only for a moral and religious People. It is wholly inadequate to the government of any other.” He was making an important point that has nothing to do with theology. Society can have extensive individual freedoms when people are socialized to mostly to make the right decisions without government coercion. If we loosen the social guardrails, as we have done, more government coercion becomes necessary to suppress anti-social behavior.
> It’s closer to “freedom to make the right choices.”
Not quite. It means, individuals have to have the freedom to make their own choices, because nobody can be trusted to know what the "right" choices are and dictate them to others.
By "a moral and religious People", John Adams did not mean that every one of those people must agree on exactly what the right thing to do is. He meant that the people have to have the concept of right and wrong as things they are supposed to discern, things outside themselves that aren't dictated by any other authority (or at least not any human one), and to understand that they have a duty to do their best to make the right choices. The problem with our society today is that that concept of "right" has been discarded; instead there is a different concept of "right" that revolves around adherence to whatever political ideology is favored by those in power.
> more government coercion becomes necessary to suppress anti-social behavior
The problem is that the government can't be trusted to do that job. That's what "freedom" means in the American context. That's why the US Constitution doesn't give the Federal government the power to do it. The fact that our government does it anyway is a bug, not a feature.
Your words amount to saying that freedom is only allowed when it's meaningless because nobody is actually exercising it in any way that matters.
Separately from that, I don't think that the original US constitution - you know, the document that explicitly protected the interests of slave owners, i.e. the vilest kind of filth - could be meaningfully said to be made for "a moral and religious People". Or, if we take that at face value, then that tells us volumes about the value of said morals and said religion, and it's deeply negative.
> Your words amount to saying that freedom is only allowed when it's meaningless because nobody is actually exercising it in any way that matters
It’s interesting that you think freedom is only “meaningful” if people actually engage in the anti-social conduct which they’re free to do. I would say the point of freedom is to eliminate the apparatus of control because you can trust nearly all people to do the right thing without it. That’s the highest form of society.
That slavery existed is not some trump card that negates everything else. It’s also a particularly uneducated comment to level at John Adams of all people. The idea that slavery is intolerable, which you easily hold in your head in 2025 without having worked for it—was bequeathed to you by John Adams and his ilk. In 1789, you would have looked the other way at slavery, just like you look the other way at everything you tolerate today. You probably would’ve even called John Adams a religious nut for believing everyone was created equal in the eyes of god, and demanded scientific proof of that.
> The idea that slavery is intolerable, which you easily hold in your head in 2025 without having worked for it—was bequeathed to you by John Adams and his ilk.
That's a rather US centric point of view. Slavery was considered intolerable in many places, the US was different in that it actually allowed it for as long as it did. Of course there are many guises for slavery that are practiced in other places but on a moral level lots of people realize it is wrong and John Adams had absolutely nothing to do with that.
Yes, I was talking about the U.S. Though I don’t know anywhere that opposed slavery with the moral fervor of Anglos. Who went to war and killed their own people to free slaves from a different ethnic or religious group?
Slavery was common in large parts of the world that all dropped it well ahead of the United States, and in most cases without a civil war (though there were some):
You could spend a good chunk of a lifetime studying this subject and still not have the complete picture. The main difference is not the 'opposition to slavery with the moral fervor of the Anglos' as much as the resistance to getting rid of slavery.
That is what sets the US apart, the stark division between the pro and the con side and the fact that the South figured out that this was the thing that they could not give up. And their roots were just the same as the side that opposed them, they just had an economic interest.
I agree that social media is exacerbating a lot of problems right now, and I don't have a ready answer as to how to fix that (or if that is even possible at all - it wouldn't be the first time a society is radically disrupted and reshaped by new tech). One thing I'm pretty confident about, though, is that heavy-handed regulation will not only not solve that problem, but will create many others. Maybe if we had some kind of widespread supermajority social consensus on this, it might have worked, but we don't.
Thanks for taking the time to answer. In my view, it's easy to legislate away at least 50% of the harm. Not perfect, but much better than 0%. This would be banning recommender systems and infinite scroll feeds for platforms above a certain size ("gatekeepers"). I'm sure there's a number of loopholes, and you might still believe that even this kind of measure would create more than it solves.
On the supermajority bit, in more and more countries there's already a supermajority for banning phones from schools, more and more for banning it for children altogether, and so on. I think that's a clear sign people would actually be in favour of quite a lot of measures to tone things down, like the one I mentioned. Total bans I wouldn't vouch for.
If true, that's definitely a US localized thing. The places I've lived in the big winners in the big lotteries are disproportionately often middle class compared to what you'd expect if the large majority of buyers were poor.
Generally, yeah, to use the online government and financial services in Sweden need BankID, which is almost always on your mobile phone. You can also use a PC, although that is fading away.
ID cards are also a thing, and in principle every grownup should always be carrying ID although its not like everybody really does when walking around the park etc.
There are paper and in-person alternatives to the online services, but the ease and prevalence of the online services makes those actually relatively efficient. The times I've had to do something in person has all been slick.
I think underneath the key concept is that everyone has a unique ID number and means to prove it's them. 99% of the time that ends up being Mobile BankID.
> Generally, yeah, to use the online government and financial services in Sweden need BankID, which is almost always on your mobile phone.
Can someone explain to me why phones seem to be considered more secure than online communication channels or desktops? The way I see it, it's a computing device you install all sorts of crap on, sourced from all sorts of questionably trustworthy sources (especially as all sort of retail companies have started moving from loyalty cards to apps).
The Estonian solution from the early 2000s - a dedicated identification device, seems far more secure and reasonable than the modern Swedish one. If any bank in my area started offering YubiKey in leu of app authentication, I'd switch to it in a heartbeat.
Because it can be more secure and everyone has one. And it can be made quite hard to tamper with, unlike your average desktop. Consider that apps are sandboxed by default, and hardware secure element key storage should be available. Of course a Yubikey would be better.
> ID cards are also a thing, and in principle every grownup should always be carrying ID although its not like everybody really does when walking around the park etc.
Tangential: I may be misinterpreting what you mean by ”should”, but no one is required by law or regulation to carry identification on them in any Nordic country (except for in certain circumstances, like while operating a vehicle that requires a license).
If the police have a valid reason to ascertain your identity as part of performing their duties, and you refuse to tell them your name, date of birth and address, or they have reason not to believe you, they can detain you until your identity has been confirmed. An id card can save you that hassle.
So if you’re just saying that it’s a good idea to always bring your id with you, then sure.
Correct! And he hit a nerve laying this logic bare, which is exactly why The Guardian took the unprecedented and absurd measure of removing his manifesto from a decades-old article when it went viral a year or 2 ago. Lest the populace get to see things for what they are. At the same time unmasking their veil that despite their pretenses, they're just as much of a capital interests-captured rag as their "competitors".
It is absolutely possible, reasonable, and normal to be unhappy with US foreign policy in the Middle East dating from before I was born, but also believe I have no power to change course there.
I'm not the "both sides are the same" kinda guy, but we realistically only have choices between Democrats and Republicans here, and I'm not convinced things would be meaningfully different in the ME if some US elections over the past 50+ years had gone the other way.
As a personal example: I was eligible to vote in my first US presidential election at age 19, in 2000. I did not vote for Bush. He still won, and still started a war with Iraq under false pretenses. I was utterly powerless to change those events.
So I'm not sure what nerve was hit here, because, "I'm going to attack regular folks in another country because I don't like what their government is doing" is still just as cruel, cold, and ineffective, regardless of the targeted country. Ultimately bin Laden failed: US foreign policy in the Middle East has not changed in ways that I expect he'd be happy with. Perhaps his consolation prize is that, due to the reaction to his attacks, the US is a lot less freer than it was before; not only did the government use it as an excuse to curtail freedoms, but the actual citizenry started prioritizing some higher measure of safety and security over freedom. But I don't think that makes the world a better place, even if bin Laden might have considered that particular aspect of things a success.
