Viewing profile — cobbzilla
cobbzilla
HN member- Joined
- Fri, May 10, 2013, 6:23 PM UTC
- HN karma
- 3,050
- Public activity
- 1,126 items
- HN profile
- View on Hacker News ↗
About cobbzilla
Recent public activity
-
comment
Comment #49185491
The proposed “feed” solution is functionally indistinguishable from incremental reconciliation. The article presents a good framing and is well-written, but doesn’t really propose …
-
comment
Comment #49162939
whatever it is, it will be heavily skewed in favor of whatever there is a lot of examples of. AI could bolt together many cookie-cutter pieces into a very large YOLO system. The tr…
-
comment
Comment #49140896
[flagged]
-
comment
Comment #49140867
> but gotta love me some Buddhist pedantry y’all can spar with the best above comment == Q.E.D.
-
comment
Comment #49120052
This “data driven” science approach sure looks like empiricism from a distance. What am I missing?
-
comment
Comment #49119209
California has codified a water caste system. In 2015 (my last experience with this), citizens were ordered to cut water use 25% while ~400 pre-1914 claimants with magical "senior …
-
comment
Comment #49106217
it’s a loose usage. necessary but not sufficient type of thing. if all your DNA instantly disintegrated, the process ends quickly, you are not you in many senses of the word. to pr…
-
comment
Comment #49106202
I use and prefer screen I use tmux on windows and elsewhere when I must They both have pluses and minuses, neither is perfect and both (yes even tmux) seem “old” in today’s world T…
-
comment
Comment #49106018
i love this. i wistfully clicked the web ring links and was sad to find they were self-refs. in another universe (or this universe in the near future?) there is a legit web ring of…
-
comment
Comment #49087083
It is in fact how Deinococcus radiodurans works: Deinococcus radiodurans has a unique quality in which it can repair both single- and double-stranded DNA. When damage is apparent t…
-
comment
Comment #49086063
I want this mutation. You have little nano-crawlers going up and down your DNA strands fixing problems. It’s like ECC RAM for the essence of YOU. And you’re basically immune to can…
- comment
-
comment
Comment #49068726
Yes, what other opex is there? It will have good ventilation, I’m not worried about cooling.
-
comment
Comment #49067930
At my house. I have 5Gbps fiber and could pay for 10 or 25 if I need it.
-
comment
Comment #49067865
I’ve priced it out: max $135/month to run a dual Xeon 2U server with 3T RAM & 2x 22 core Xeon Gold. It’s the 2x 750W power supplies that ultimately determine opex. My power costs $…
-
comment
Comment #49051485
311 may have said it best ~33 years ago: I tell you what's wrong to me The industry's ability to manufacture stars They churn them out like Iwo Jima makes cars But the blame lies n…
-
comment
Comment #49046063
If you mean being forced to overpay for no-longer-SOTA closed models, that’s I guess the protectionist motive here: impossible to enforce, enforcement uneven and political, illegal…
-
comment
Comment #49044676
How do you do that without banning the model itself. It doesn't matter who can or can't run it. They are trying to ban something that can be represented entirely by a sequence of b…
-
comment
Comment #49036171
The AI rigs cost money, sure. But no one is talking about regulating those. We are talking about regulating models, which do have zero cost of reproduction. It’s like banning the l…
-
comment
Comment #49030405
The first token on a new AI rig costs $X (full capex cost) then every token after that costs virtually zero. Over time the cost/token trends towards zero (modulo opex). That said, …
-
comment
Comment #49027464
I can understand (not necessarily defend) protectionism for products with nonzero manufacturing cost and some required physical size and such. I am befuddled when the same policies…
-
comment
Comment #49019938
No, the court case you’re thinking of was when someone tried to say the AI itself had the copyright ownership. That’s not allowed (at least in the US), the copyright owner must alw…
-
comment
Comment #49016656
If you’re processing N things at a time and your “scalar tail” is N-1 why can’t you put in a dummy value for the last entry, run one last SIMD iteration, and discard the dummy retu…
-
comment
Comment #48980241
Very good point. It’s nontrivial but not impossible to determine this, I am very curious! One nightmare scenario is a 3-bit error being detected as a correctable 1-bit error upstre…
-
comment
Comment #48968570
> they could just use a regex to change Claude to Kimi in their distillation pipeline! Jean-Kimi Van Damme would like to have a word with you.