Rodmine

AI is great for one-shotting impressive demos and fixing well-defined things; it's not even close when it comes to very long context and really complex tasks. If you haven't created a harness around your complex application, the use of AI will hinder rather than help your project.

I am showing my bias openly. Glad you were biased in that direction to notice.

Yes, in the expected range but could have been worse so happy about it.

Who do you imagine I insulted? If you get something for $2 but you pay $70 what can I say about you that is nice?

Same... but then it's exactly what you'd expect from Europe: mediocre, overpriced, unclear license, expensive to operate, but could have been worse. It's not as horrible as you'd expect Europe to be.

At this point, are you using Deepseek flash 4.1 or is somebody else paying for your model usage unconstrained or are you a moron?

It reveals what they are to the public. It's fake crypto with all the downsides of crypto but none of the upsides.

Love the commitment by “Amishpunks”. I am all-in AI now but can imagine a future where I will be one of you guys.

You have to be a total imbecile to think that we are capable of collecting enough data to make such conclusions.

During a gold rush, sell shovels. They are investing heavily in in-house custom silicon, which is the shovel end of things. They could open more jewellery chains, but that is a very different kind of operation and it is in their interest to let others compete for that and use their resources on the shovel side of things.

Not surprisingly they also sucked some US dicks during the 5G ban on Huawei back in the day: https://www.bbc.com/news/technology-45281495

This is not about merely running it for demo purposes; it’s about running things in production for generating actual value.

If the CCP was so advanced that they can implement these attacks in models released as .safetensors, I would assume that the government can pay multiple western frontier labs can show how and where this is done and publish it for anyone to reproduce in a sandbox.

In reality, the attack is much simpler: flood the market with genuine competition so that neutral parties (and their own corporations and that of their allies and potential allies) never have to be enslaved by the US corporations and hence they cannot be forced to join the US in sanctioning China.

This is much less risky as a strategy (imagine how bad getting caught doing the other thing would be) and works way better.

I think it was never OK to enjoy programming, at least if you were working professionally. If you enjoyed programming, I would always suspect that you were probably creating programs that were not needed, adding complexity, technical debt, security issues and long-term liability, for the sake of your personal enjoyment.

So someone should pay them their hard earned money for them to not have to solve problems that are actually relevant? They should never grow up to learn how things actually work and make contributions that actually matter?

For those who missed the point: this post is related to the Deepseek ban. You can't have open source AI models, because the "distilleries" don't like it.

It is THE competitor right now when it comes to price and intelligence. Also, just banning it plants fear in other sponsors and investors. When they do it like this, they can pretend that this is not a part of regulatory capture but it is an actual security concern because "this one in particular is dangerous" and "we'll tell you when others become dangerous" which will happen when they become a competition (now less likely because investors are cautious).

No, it is not.

I feel like the people you are talking about in the latter just want life to be like when they were in 8th grade. You go to school and your parents and teachers do everything for you.

Welcome to the real world, where you have to solve real problems.

Well, it's a bigger improvements in Deepseek 4.1 Flash and Deepseek 4.1 Flash 1017 than whatever Google was doing between Gemini 3.1 and Gemini 3.8.

It was confirmed that it was not the same software that they provided the source for back in the day: https://cybersec.ee/2018/02/20/reproducing-vote-verification...

oH tHeY mUsT hAvE mAdE sOmE lAsT miNuTe bUgFiXeS.

Everyone with half a brain should know that this is completely compromised. Nobody wanting to run fair elections would support electronic voting.

The country which sold "climate change" to the world to try to push their "clean-tech" energy market products as a defiance of OPEC embargo, is now trying to push their "care" for fish to discourage hydropower.

True, it isn't neck-and-neck. It's neck and balls. Deepseek's balls and Opus 5.5's neck. Whatever you can build with Opus 5.5, I can build in the same time, 1/500th the price with Deepseek Flash 4.1.

No, it's open source and is actually cheap to run, while performing the same, which is why it is being run and provided by 31 different providers for Openrouter.

Here is a benchmark where they actually evaluated harnesses across same models and measured task completion rates: https://frontierharness.org/

He most definitely had a hand in murder of the 26 year old whistleblower: https://www.youtube.com/shorts/lL7WA6i3zVw

Mother's interview: https://www.youtube.com/watch?v=Kev_-HyuI9Y

Google definitely does not have good models and if it does, Gemini Flash series models are definitely not them. I use the Google Gemini app daily (for searches, research etc.), everyday, for some query, gemini flash 3.8 hallucinates, does not understand context etc. in quite short conversations. I switch back to 3.1 pro and it works fine.

Which brings me to why do they get good benchmark scores? Have never really understood this. They must be optimizing just for this.

3.7-flash has been useless many times, specially when context gets bigger. 3.1 is the only Google model that has seen use from me. With extended thinking, 3.7-flash is kinda usable but not without many problems. I find myself falling back to 3.1 often. I don't believe in any benchmarks because whatever they are doing to award 85% to 3.7 on anything, they should seriously reconsider that test for anything.

Check your diet. ADD/ADHD has been linked to histamine intolerance.

He didn't commit suicide. He was most definitely suicided.