That has nothing to do with context window or the models at all; it's the compute capacity.
Pretty much every single AI firm had the same issue when the new model came out, it's the result of everyone jumping onboard and consuming all available compute capacity to break it for days.
Claude had the same issues when they released their newer models.
A lot of folks are finding GLM 5.2, Kimi 3, and Deepseek just fine for their use cases.
It does not have to beat US firms, it just needs to be cheaper.
I use Deepseek v4 flash for a lot of reviewing and summarizing tasks, only used 4$ in the last two months. No dramatic drop in performance against other US models, it works for my use case. I do use GPT 5.6 Sol for other things but tried GLM 5.2 and it was good enough.
Just an update, because I think my last point wasn't clear.
I'm not saying forever drug is a bad thing. I'm saying people who wants to take GLP-1 as a temp drug to lose weight quickly and get off it may not realize the long term impact it may have on their body.
For an example, certain type of anti-depressants has serious side effect that if not carefully tapered off and managed with a different drug, can leave folks with cluster headaches that lasted several months if not years.
Nobody understands why it all helps, they just noticed it does work for something, quickly pushed it to the market first to get patents and get all the profits from it before generics + derivatives hit their pockets. Now, everyone is studying it because there's all new funding coming in for it and finding other versions of it that they can profit off it.
Nobody knows what migraine really is, so this isn't a surprise to them that GLP-1 may help, the main question is; why? So they have another data point proving that gut health has a direct correlation to the brain.
Keep in mind that a lot of the benefits go away once patients come off GLP-1 and we have not seen any studies yet on what happens to people who come off it for long term effects. It may in fact make things even worse and for a lot of people, they may have to stay on it for the rest of their lives.
Yes. Gut microbes has already been shown to have a great impact on how we metabolize by what med we take, what we eat or drink and intake from our environments (micro-plastics, etc).
There is no single main root cause for obesity. We just combine it as one because there isn’t a lot of long term research or funding for it right now. There is a lot of sigma against obesity and people keep blaming other people instead.
Thyroid hormone disorders have been linked to cause weight gains. This can’t be fixed by simply eating less, it can literally do far more damage.
Medications have been linked to cause weight gain as side effects. This wouldn’t do anything to eat less until they stop taking meds and for some, they cannot do that.
Americans’ increasing desire for sweets have increased the sugar content in all of our food including the fruits and vegetables over time. We’ve intentionally bred our healthy stuff to be sweeter. So eating less can make us even more hungrier because we go into sugar crush without realizing it. Changing diets is difficult without us doing all sorts of calculations of finding the right cheap healthy food at the right store and that is you are lucky enough to have any.
Can you show me what we're doing in USA to help children and people develop the habits and discipline for long term lifestyle change?
Because I've never learned anything about nutrition, macros, high sugar content and all of the healthy food I should learn to eat on my own.
We did not have home classes in any of my education in US at all, they were a thing in the past but that wasn't a thing in my middle hs or hs or college at all in NY in 90s/2000s.
All of my bad habits were from my parents and they were not good eaters.
Great news app. Not only does it learn which news you'd like to follow, its AI feature is actaully useful for a change (for me at least); it extracts the data you want to know about and it has a few useful views such as the "opposite views" primpt that I really like using to learn from both sides.
Related articles are great to reduce duplicates and being able to ask questions or seeing questions from other users was great to see.
That’s not what the OP said though, what you’re describing is basically regulations. All we can do is regulate to outlaw and punish those who don’t comply with it but that’s still not possible to limit.
Chemical weapons are still being developed (Russia used it in their war against Ukraine), bioweapons are still being developed, meth are still being manufactured all over in US if not smuggled in.
We can’t prevent the technology from being invented, it was always going to happen (video games and VR was practically going to have this AI generated worlds).
It is just a matter of doing it right the first time such as having license/agreements or company building it based on their own images.
A lot of stores don't accept them anymore because of scumbag scammers abusing the system, leaving us dependent on a friend or family member to make our calls.
I don't know if you're outside of US that's done differently but in US, it's not called "deaf relay", it's just relay services.
The relay service is not for deaf people, it's also for people with speech issues, deafblind, blind, and so on.
> transcriptionists.
Also, not as such in US. It's interpreter (usually for ASL/video call) or operator. Relay service relays messages in two ways; voice the message from the end user and then type from voice to the end user.
AI has to be absolutely perfect to replace humans for these relay calls because as you said, when it comes to medical situations, or legal issues; who would be at fault for miscommunication that led to some issues?
In US, you must also verify your identity to the relay service as federal regulations has step up the identity requirements recently. I had to do a video call with VRS to show my state ID that I'm in the state and registered with the state's relay services.
The issue was more specific to higher token burn rates, not latency.