Viewing profile — amitport
amitport
HN member- Joined
- Sat, Jan 25, 2014, 5:45 PM UTC
- HN karma
- 1,339
- Public activity
- 569 items
- HN profile
- View on Hacker News ↗
About amitport
Recent public activity
-
comment
Comment #49229910
I've been considering something like this for a while. It only makes sense as a free, open-source, decentralized sharing protocol (where anyone can host theorems and no one can lim…
-
comment
Comment #48599968
C# did not ship with async/await, and Java didn't have virtual threads back then. I am specifically referring to the initial choices made in C#'s foundation.
-
comment
Comment #48596614
The mystery of why .NET got so many things right is simply that C# was built several years later by the exact same Microsoft engineers who had previously worked on extending Java, …
-
comment
Comment #48525969
[dead]
- comment
-
comment
Comment #48525682
Let’s look at the actual historical record, because the pattern is undeniable: the root cause of this conflict is the absolute refusal of the regional and local leadership to ever …
-
comment
Comment #48525103
"parents and grandparents," You don't have to go into historical events. This is still happening now. Jews are still fighting for their survival and the moment Israel stops fightin…
- comment
-
comment
Comment #48488150
[dead]
-
comment
Comment #48470820
To be fair, NPM sucked long before it got acquired by Github/Microsoft. And to be fair 2: The other package repos also suck.
-
comment
Comment #48462989
[flagged]
-
comment
Comment #48146203
Hi, thanks! I appreciate your input and generally agree. The TDS article wasn't really aimed at the HN crowd, but it did help a bit with the more general audience. I do plan to als…
-
comment
Comment #47998505
For context: https://towardsdatascience.com/how-a-2021-quantization-algor...
- story
- story
-
comment
Comment #47987284
I recently wrote a beginner-friendly explanation of this situation in TDS: https://towardsdatascience.com/how-a-2021-quantization-algor...
-
comment
Comment #47986318
Hi, I’m the author of the post above. tl;dr TurboQuant is a recent paper from Google and NYU that has gained massive traction in mainstream media and the AI community. As implement…
- story
-
comment
Comment #47922845
In the vLLM documentation quoted above, TurboQuant (which is a restricted version of EDEN) is referred to as a specific case of HIGGS. Note the symmetry: EDEN acts as a special cas…
-
comment
Comment #47922731
Thanks for the pushback, and I appreciate the reference to classical information theory. While I probably overstated things by using the very general phrase "taking advantage," I w…
- comment
-
comment
Comment #47921796
Thanks for that! It is worth noting that taking advantage of the post-rotation distribution was not actually done until DRIVE (2021), which was made possible via our proper scaling…
-
comment
Comment #47921299
Thanks for that! Note that the residual chain is empirically and theoretically inferior to our unbiased scale; furthermore, it requires an additional bit in certain cases. Addition…
-
comment
Comment #47920896
Those works did cite DRIVE/EDEN :) HIGGS is an extension of EDEN (using the well known method for blockwise Lloyd-Max). The proper framing of this "TurboQuant" layer in vllm (which…
-
comment
Comment #47920259
I believe our claim at this point is more fundamental than just lack of citation. The quantizer in TurboQuant is EDEN quantization (2021) applied to the KV-cache. It is neither a n…