Live data from Hacker News

Viewing profile — augment_me

augment_me

HN member
Joined
Mon, Mar 31, 2025, 8:19 AM UTC
HN karma
424
Public activity
102 items

About augment_me

No profile information was provided.

Recent public activity

  1. comment
    Comment #49158570

    Alternative title: "Our invention makes INT8 better on older consumer cards without FP8 support"

  2. comment
    Comment #49135920

    As some other commenters pointed out, the comment (and the book) are not making claims about work hours. Seeing it from the perspective of "dawk to dusk" with our, now time-based p…

  3. comment
    Comment #49131072

    How bout a really big iceberg chunk, like massive. There has to be a trade-off at some size. Also nice reference and funny in probably most ways to think about it thanks for a laug…

  4. comment
    Comment #49130896

    In "Four thousand weeks" by Oliver Burkeman, he breaks down that this obsession with action stems to the industrial revolution when it was decided that workers should sell their ti…

  5. comment
    Comment #49118806

    Who cares there is so much more to exploit. For example no-one has yet moved massive freshwater icebergs by boat, and its up for grabs!

  6. comment
    Comment #49092750

    I feel like the Kimi team is amongst the best in the industry to pick and choose what is meaningful from the other models. For example, avoiding the expensive and empirically uncer…

  7. comment
    Comment #49092695

    When you have recurrent blocks in your model, you implicitly have a timestep T(amount of recurrent steps). Similar to Diffusion Transformers, it then becomes valuable to encode the…

  8. comment
    Comment #48923059

    Just one more harness bro, one more harness I swear one more harness and we will have solved AGI bro

  9. comment
    Comment #48744071

    Peer review WAS vital for a long time. Maybe the world looks different now, maybe LLMs can find value in things better than humans. When you make an assumption it's good to think a…

  10. comment
    Comment #48744050

    You are then our of the normal probability distribution and out of luck, it's not profitable to cater for you for the company.

  11. comment
    Comment #48731593

    My comment is aimed to highlight that the "GPU Bubble" is frames as a general solution when it's not, its a specific bottleneck based on your model size. Your dont mention your mod…

  12. comment
    Comment #48729928

    As someone who works in the field, the blog is nice but it has a lot of CODEX fingerprints on it, and it's also very specific to the size of the model in question in a way that is …

  13. comment
    Comment #48724406

    Whether brains do gradient descent is irrelevant to a CFO deciding whether to staff one radiologist or three. The market doesn't care if the model "identifies" versus "returns a st…

  14. comment
    Comment #48711105

    You could also view it from the perspective of that if every other major superpower has their mass surveillance and you don't, it becomes an assymetrical informational situation wh…

  15. comment
    Comment #48683084

    If you have a confounding variable or a dependency that influences the experiment to a degree that invalidates the premise of the experiment, you need to put more weight on this in…

  16. comment
    Comment #48683037

    LLMs produce about 95% of the code at my company and review about 70% of it for 3 years now. Our team has downsized from 40 to 8 people in this time. My creative labor is spent wri…

  17. comment
    Comment #48682938

    Both were noted, but then the conclusion drawn from these things is that the author is considerably more optimistic about the agents. In my opinion, if you have factors that narrow…

  18. comment
    Comment #48682896

    Who cares if the consumer buys it and uses it? Information is worth nothing anymore, attention is, so if they manage to capture a larger audience somehow, they win.

  19. comment
    Comment #48682880

    1) Googles spam filter removed a lot of the attempts as you say yourself. 2) Model was tested under unrealistic conditions where 99% of the inputs are malicious, so the model is ex…

  20. comment
    Comment #48669952

    Which is interesting given the Mojo blogs where they shit on the other pythonic eDSLs like Triton saying that it's a dead end

  21. comment
    Comment #48633068

    Bitter truth :(

  22. comment
    Comment #48633040

    Most people use social media such as discord or whatsapp in order to make social activities and communities simple with the majority of their friends. A majority of people do not g…

  23. comment
    Comment #48347234

    > The correct solution is just to not attend it if you know you aren't requested to participate and are just here to grow the numbers and make your company waste money. If this arg…

  24. comment
    Comment #48346803

    I feel like the whole blog and the point can be reversed. If your bottlenecks are meetings and emails, and you make an agent take notes and summarize things for you, you gained foc…

  25. comment
    Comment #48233761

    TLDR: Authors realize that global row-wise dependent functions like RMSNorm/LayerNorm have baked-in scales that are commutative in certain setups, so they can be moved out after a …