- They had all these benchmarks where they were saying Mojo is 68,000x faster than Python or whatever. Some of these blogs/benchmarks got deleted because they were awful, comparing the worst possible Python code to the most optimized Mojo code.
- The same with "Mojo is faster than cuBLAS", it means it's faster in a particular benchmark in a very particular case, and it's actually slower in other cases. But the headline is what matters.
- A superset of Python, it will probably never happen. You will never see Python/Mojo like C/C++. A fake promise.
Apparently, it's what happens when you get Silicon Valley money, you start being like this.
I asked z.ai, "Which is the best model for coding"
Here's the response:
```
As of late 2024, there isn't one single "best" model for every situation, as performance depends on whether you need raw coding intelligence, speed, or integration into your workflow.
However, the current consensus among developers places Claude 3.5 Sonnet at the top for pure coding ability.
Here is a breakdown of the best models for coding right now, categorized by their strengths:
...
```
So they have data until late 2024 and nothing beyond? They don't even perform a web search. Doesn't seem to be on par with other frontier models.
reply