I like their models, super cheap - I'm a Lite plan subscriber, and subjective performance seems to be same as lower Anthropic models, useful for lots of grunt work.
The problem is that Ziphu really __really__ struggle with capacity - everyone is complaining of timeouts or very slow speeds. I can't get direct access to the model though I see it is in OpenRouter so I may play. But the capacity issues means DeepSeek is my main provider these days
Yes, precisely. Assessing your own cognitive skills is dubious. I’m pretty certain I’m less clever than I was when younger but if I find a problem tough now maybe 25 yo me would also have struggled?
I watch it and think of the “Are we the bad guys?” meme: lots of people advancing their careers in the Empire & shrugging at their own evil deeds mentally saying “ ok, but what can I do?”
Not saying you are wrong about many US business owners but having travelled in China I can tell you that Chinese companies have no qualms about poisoning their neighbours for enhanced profits - my guide there said it was what he hated about his country. Money drives ethics pretty much everywhere I reckon.
In a previous life I've employed contractors and software engineers to run a criminal website. Motivations for my guys were that it was well paid work that was technically challenging in order to evade enforcement agencies, and was 'fun' in that respect; they were "sticking it to than man (my service was regarded as moral by all my users & others); and there wasn't so much work about that they could pick and choose; lastly, I was a good employer because I had to be!!
His "A Declaration of the Independence of Cyberspace" was the most unintentionally comical thing I read at the time. Like his lyrics and prose it was woefully pretentious & leaden.
I also met him once. A more unpleasant, up his own ass, person I can barely recall.
I tried getting any sota llm (GPT 5, Opus 4.6, Deepseek V4 pro, glm-5) to write a Metal 4 shader for a bottle usdz and none of them got it right. They screwed up the normals and textures , total mess. I tried it to do it in Metal 3 and still crappy.
I also remember being amazed, and did a forehead slap, when an old army bomb disposal man explained how, what I thought was an innocent device, was used by the IRA in bombs.
I did the same using the mlx version on an M1 Macbook using LMStudio integrated into XCode. I had to up the context size I ran it a against a very modest iOS codebase and it didn't do well, just petered out at one point. Odd. Pretty good chatbot and maybe against other code it'll work but not useful with XCode for me
Get down to A&E quick, you've clearly drunk a potentially fatal amount of Elon KoolAid.
Musk is a buffoon. Clever? yes by all accounts, genius? Hardly. He's had luck, made good judgments mostly offsetting the bad ones. Most of all he has enough money to power through errors that would bankrupt thee & me.
When I was a boy into electronics the one thing that always stymied me designing circuits was how to make an inductor of the right value. I can't recall these many decades why but I needed to make ones of a value and you couldn't specify "xuH" to a component supplier, like resistors or capacitors, but nor was there any resource to say "x turns of y gauge wire" etc. It was all a black art to a 12 year old in the early 70s.
Similar experience to me. I tend to let glm-4.7 have a go at the problem then if it keeps having to try I'll switch to Sonnet or Opus to solve it. Glm is good for the low hanging fruit and planning