I usually ask a simple question that ALL the models get wrong: List of mayor of my city [Londrina]. ALL the models (offine) get wrong. And I mean, all the models. The best that I could, it's o3 I believe, saying it couldn't give a good answer for that, and told to access the city website.
Gemini 3 somehow is able to give a list of mayors, including details on who got impeached, etc.
This should be a simple answer, because all the data is on wikipedia, that certainly the models are trained on, but somehow most models don't manage to give that answer right, because... it's just a irrelevant city in a huge dataset.
But somehow, Gemini 3 did it.
Edit: Just asked "Cool places to visit in Londrina" (In portuguese), and it was also 99% right, unlike other models, who just create stuff. The only thing wrong here, it mentioned sakuras in a lake... Maybe it confused with Brazilian ipês, which are similar, and indeed the city it's full of them.
Ha, I just did the same with my hometown (Guaiba, RS), a city that is 1/6th of Londrina, and its wikipedia page in English hasn't been updated in years, and still has the wrong mayor (!).
Gemini 3 nailed on the first try, included political affiliation, and added some context on who they competed with and won over in each of the last 3 elections. And I just did a fun application with AI Studio, and it worked on first shot. Pretty impressive.
(disclaimer: Googler, but no affiliation with Gemini team)
Pure fact-based, niche questions like that aren't really the focus of most providers any more from what I've heard, since they can be solved more reliably by integrating search tools (and all providers now have search).
I wouldn't be surprised if the smallest models can answer fewer such (fact-only) questions over time offline as they distill/focus them more thoroughly on logic etc.
Funny, I just asked "Ask Brave", which uses a cheap LLM connected directly to its search engine, and it got it right without any issues.
It shows once again that for common searches, (indexed) data is the king, and that's where I expect that even a simple LLM directly connected to a huge indexed dataset would win against much more sophisticated LLMs that have to use agents for searching.
If is anything like the WhatsApp Web, "works" might be optimistic choice of word.
When I switched from Windows, the thing that I missed from Windows on Linux was the native WhatsApp App. Now they killed there, so feeling better on my switch now!
The WhatsApp web app is not perfect (no software is) but I’ve never had any major issues with it. It’s snappy and very rarely does it glitch out. I find your comment surprising.
It's a ram hog if you stay the tab open for a while, with ton of messages. So from time to time I need to close the tab and open again.
Huge perf issues because of this.
Also had some serious bugs for a few week. Had to let WhatsApp Web wait for completely sync for 15 minutes~ or else it just stop responding and crashes everything.
Windows also need this for 10 years, since Windows 10 launch. But never happened. I guess it's just not really viable, as they need to continue selling new devices...
I still don't understand how there can be like 10 different UI styles in Windows 11 with no progress to getting that fixed. It feels more likey they are adding new ones instead...
Especially considering how much software these days on Windows are all Electron/Web. So is not a hard switch as it once was.
I switched from Windows to Linux it's been 2 years. One of the few things I missed on Windows, was the native WhatsApp app, as the Web WhatsApp it's horrible. Then a few months Meta killed the native app and made into a webview-app :)
It only takes one application to force you back to using Windows.
e.g. HellDivers 2 didn't work well until recently on Linux. If you are playing certain factions it is a very fast paced game and I would frequently experience slow downs on Linux.
So if I wanted to play HellDivers 2, I would have to reboot into Windows. Since running kernel 6.16 and updates to proton it now runs better.
It's bad too, yes. But not as bad, because MS is a profitable company with real enterprise products, so they have some reputation and compliance to maintain. SamAI is a deeply unprofitable company, mostly B2C oriented, with no other products to fall back to except for LLM. So it is more probably that Sam will be exploiting user data. But in general both are bad, that's why people need to use Firefox, but never actually do so, due to some incorrect misconception from decade ago.
>MS is a profitable company with real enterprise products, so they have some reputation and compliance to maintain.
On the contrary, it could be the case that Microsoft ritually sacrifices a dozen babies each day in their offices and it would still be used because office.
In Brazil, medicines lose patent protection 20 years after the original filing date.
The filing date was in 2006, so the patent expires in 2026.
But note that the patent was only granted in 2019, it took 13 years. They went to court last year, but the justice (as far as I know) has continued to follow the precedent that the original filing date applies.
So they went to the federal government to try to change the law, but the government has refused so far.
The INPI (Brazilian Patent Office) used to take a VERY long time to register/grant patents. It's faster now (about 4 years on average), but it's still slow.
But I'm seeing here that in the US it was 2017. So not that different from 2019 anyway.
PS: Brazil also have several local labs focused on generics, and besides this, there's also state-owned lab, Fiocruz, who makes vaccines and medicines as well, several of them, to distribute though SUS.
In Brazil by the constitution, everyone has the right of health, so if the public health care (SUS) don't distribute, you can sue the governament and get the medicine. So there's a big incentive for the governament to make patents expire as soon as possible to have generics and include on SUS.
e.g Fiocruz manufacture PrEP, and started now PrEP injection (Cabotegravir).
In Brazil it will expire in July 2026. It's pretty relevant as it's kinda already announced they will put the generics on the public health care (SUS) for free... Which is big deal as an Ozempic shot costs almost the same as the minimum wage.
This year they already did an analysis to include Ozempic, but it was denied, probably because of the cost difference...
But they were trying on justice to extend the patent...
SUS doesn't have a good track record when it comes to offering modern drugs and treatments. It takes years if not decades for medications to be covered, even when generics are available. Patients often had to go to court with documents from their doctor in order to make things happen. I've lost count of how many patients I've helped get their treatments... And now the supreme court has closed the door on them.
Still, after dapagliflozin I'm ready to believe it. There's also the fact semaglutide will straight up win votes for politicians that provide it.
I expect it to mirror the evolution of dapagliflozin in SUS: it will be offered unconditionally at first, then as they start to feel the weight of the costs involved they'll start imposing conditions and forcing patients to go through endless bureaucracy to get their medication. Age requirements, comorbity requirements, obnoxious paperwork that "justifies" the need for such a drug. Oh I'll gladly fill out those forms...
Depends the price of the generics, of course. But I wouldn't doubt if the governament itself would manufacture it. Fiocruz (federal gov lab) already manufacture several drugs.
Also, Rio mayor already said he will just give to everyone basically lol
You also need to remind that on a public health care system, having less obese people means you'll spend less on health care.
being free means taxpayers will pay. and being on the public system means it will be prescribed only for health risk, just like plastic surgery already is. you don't see people getting nose jobs just because.
Generics are so cheap and easy to make that they can reasonably justify that it will save the government money by making the population healthier. Just think about how much less heart disease and cancer and every other obesity-affected disease costs the taxpayer.
totally agree. was pointing out things left out from the previous comment which was kinda of impling gov would be paying all uses of the sliming drugs, even if on generics price. ... because those drugs make profit not because of it's health benefits.
First, it seems obvious to me that they'll prescribe it to everyone who, supposedly, undergoes bariatric surgery through the SUS. Generic Ozempic is certainly cheaper than bariatric surgery.
That said, I believe it will go even further.
To get medication through the SUS, you only need a prescription from a doctor. Any doctor. So even if the SUS itself don't prescribe, you can just find a doctor who will do it...
But, likely SUS will prescribe anyway. Rio mayor:
"Rio will be a city where there will be no more chubby people; everyone will be taking Ozempic at family clinics." "Said the mayor, adding that he has already taken the medication. He claimed to have lost 30kg with the medication."
> To get medication through the SUS, you only need a prescription from a doctor.
Not in all cases. They can and will make the patient and multiple doctors fill out half a dozen forms to "justify" the use of an expensive medication.
Probably won't happen at first, politicians need to make good on their promises after all. But it will happen. I just got done filling out the forms for a COPD medication that's much cheaper than semaglutide. Only political intervention can save it.
> Not in all cases. They can and will make the patient and multiple doctors fill out half a dozen forms to "justify" the use of an expensive medication.
But any doctor can do it, not just SUS doctors. You know the country we live in. Private doctors will do the prescription... Actually, to buy it you already need a prescription, and yet, even skinny healthy folks buy ozempic.
And to get in the SUS, it's pretty simple here in Paraná at least (and as a SUS user). The worst that happens, it's that for expensive/controlled medicines, they centralize into a single public pharmacy in the city center, and you need to schedule a day/time, and bring the ID - don't even need to be the user, you can ask anyone you want to get it if the prescription and ID it's there.
Which is why for my cheap medicine, I prefer to buy it. Not worth for me going to city center only for 30 pills that cost R$ 15.
It is prescribed to the morbidly obese. It's just that patients almost always opt for cheaper options when I inform them of the costs of the treatment. Even if our public health care system starts offering semaglutide, it is extremely likely that it will impose requirements and require "justification" for the use of the medicine. You know, the usual soulless kafkaesque bureaucracy. It can frustrate people so much they give up. I've seen it.
From context it was clear to me that they meant the entire income.
Minimum wage in brazil is 280/month. I don't know how much of it would be taken by taxes, but even if it's 0 then ozempic in Brazil still costs almost 70% of that amount each month.
It’s not so straightforward, though. In the U.S., wages are usually discussed in terms of hourly pay or annual salary, whereas in Brazil, they’re typically expressed as monthly earnings.
Gemini 3 somehow is able to give a list of mayors, including details on who got impeached, etc.
This should be a simple answer, because all the data is on wikipedia, that certainly the models are trained on, but somehow most models don't manage to give that answer right, because... it's just a irrelevant city in a huge dataset.
But somehow, Gemini 3 did it.
Edit: Just asked "Cool places to visit in Londrina" (In portuguese), and it was also 99% right, unlike other models, who just create stuff. The only thing wrong here, it mentioned sakuras in a lake... Maybe it confused with Brazilian ipês, which are similar, and indeed the city it's full of them.
It seems to have a visual understanding, imo.