almostgotcaught

> I think you are missing the nuance between the different aspects of using the Python Interpreter, and integrating new functions with Python

What if I told you I used to work at modular? What would you say then to this accusation that I'm "missing the nuance"?

The rest of this is AI crap.

> your only recourse is reading code

Oh the horror!

> Yes. Is it using Python as the language?

You're completely wrong here. That's the "what's wrong with it".

> Mojo is a variant of python.

Lol this is how I know no one that pushes mojo on hn has actually ever used mojo.

The solution is to look at a lot of examples

https://www.hakank.org/

I don't understand what you're saying. We're not talking about the market exploiting labor because before you are hired by the company you're not labor for the company. Is this really that difficult to understand?

> This, I cannot believe how all the most pro-workers rights people I know also support "open borders"-like philosophies.

You ever consider that it's because those people are pro-workers everywhere and not just workers nearby? So yes enabling foreign workers to improve their lives by coming here makes perfect sense.

> What do you think is going to happen to your bargaining power as an employee when your employer has an infinite workforce to draw from?

I mean that's like saying "what do you think is gonna happen to your rights once all the slaves are free". The answer hinges on whether we continue to operate under the government that's comfortable with exploiting its citizens.

It just means next gen

It's the same exact thing - if some company makes you jump through hoops to get hired that you find distasteful just don't apply to company.

> Should be illegal honestly.

I can't imagine this kind of entitlement. If you don't want to work for them, don't study leetcode. If you want to work for them (and get paid tons of money), study leetcode. This isn't a difficult aristotelian ethics/morals question.

ITT: lots of people who think they understand the expression problem but don't actually understand the expression problem.

> I've used C++ for so long and I'm a good way into thinking that the language is just over

"no one goes there anymore it's too crowded"

These are my favorite types of comments on hn

Anthropic is a Jax shop

Yes and?

> not sure what's the usage of torch.compile

consider that at minimum both FB and OAI themselves definitely make heavy use of the Triton backend in PyTorch.

Triton is a backend for PyTorch. Lately it is the backend. So it's definitely double digits percentage if not over 50%.

> which are never, ever written in Python

nah never ever ever ever ever ... except

https://github.com/FlagOpen/FlagGems

https://github.com/linkedin/Liger-Kernel

https://github.com/meta-pytorch/applied-ai

https://github.com/gpu-mode/triton-index

https://github.com/AlibabaPAI/FLASHNN

https://github.com/unslothai/unsloth

the number of people commenting on this stuff that don't know what they're actually talking about grows by leaps and bounds every day...

Hmm I wonder though if the most canonical was to see who contributes to LLVM is the git blame or the list of presenters at a tenuously related dev conf. Hmmmmmmmmmm.

I see you comment here all the time about LLVM and I'm pretty sure you have no idea how LLVM works (technically or organizationally).

Yes there's no roadmap but it's flat out wrong that most contributions are from university students because hardly any contributions are from university students. It's actually an issue because we should be more supportive of student contributors! You can literally look at the contributors on GitHub and see that probably the top 100 are professionals contributing as part of their dayjob:

https://github.com/llvm/llvm-project/graphs/contributors

I don't care if it is different or isn't - I'm just saying it's completely transparent and obvious and hn is basically falling (again) for content marketing.

> Claude Code was used to automate reconnaissance, harvesting victims’ credentials, and penetrating networks. Claude was allowed to make both tactical and strategic decisions, such as deciding which data to exfiltrate, and how to craft psychologically targeted extortion demands. Claude analyzed the exfiltrated financial data to determine appropriate ransom amounts, and generated visually alarming ransom notes that were displayed on victim machines.

y'all realize they're bragging about this right?

> What’s truly brutal is acquiescing to children droning away their youth mindlessly clicking meaningless games or watching vacuous streaming content

Both of these things are droning. One is just culturally acceptable. Don't think so? Stop by your local Kumon center some time.

> This is the age of social media. This person has hit the front page of HN twice now. That's a commercially valuable skill.

In general yes, wrt HN it's not; literally in this second post he bemoans that the first one didn't pay off for him.

Because it's cringe. I'll never understand why people publish essentially personal diary entries in public spaces. It must be some kind of shame kink.

The post reads to me like all those movies about the nerd with a heart of gold that the hot girl will recognize and eventually marry.... which only happens in those movies.

Do people really not understand that companies don't care one whit about your personality? They only care about whether you can make them more money. And that extends to interviewers; the number one thing interviewers care about is can you meaningfully contribute to the existing roadmap, not whether you can bring your own unique perspective. This is especially true at mega huge corporate places like anthropic.

Almost all performant kernels employ structured sparsity

ranking models benefit from gpu impls of sort but yup they're not nearly as common/important as spmm/spmv

this is missing the most important one (in today's world): extracting non-zero elements from a sparse vector/matrix

https://developer.nvidia.com/gpugems/gpugems3/part-vi-gpu-co...

This is already the standard stack based DFS?

  def dfs(graph, source):
    n = len(graph)
    visited = set()
    stack = [source]
    
    while stack:
      node = stack.pop()
      if node in visited:
        continue 
      visited.add(node)
      for nbr in graph[node]:
        stack.append(nbr)

So I don't know what all the confusion is about...