Viewing profile — dist-epoch
dist-epoch
HN member- Joined
- Wed, Mar 22, 2023, 5:02 PM UTC
- HN karma
- 2,625
- Public activity
- 2,256 items
- HN profile
- View on Hacker News ↗
About dist-epoch
No profile information was provided.
Recent public activity
-
comment
Comment #49221117
Military has a phrase for the outcome - collateral damage. > Yes, I just hacked into AWS and shut down all of the data-centers, because it's where Anthropic Mythos servers are host…
-
comment
Comment #49221102
OpenAI reported the Artifactory vulnerability, patched it, then the agents immediately found a new zero day.
-
comment
Comment #49221085
> If anything, I want these models to be less persistent at their focus of completing their goal, and instead just call defeat and say “I’m not sure how to proceed next” This goes …
-
comment
Comment #49216393
CXMT is sold out too. https://www.guru3d.com/story/cxmt-dram-capacity-reportedly-s...
-
comment
Comment #49214161
Oracle is not a software company, but a litigation company. It will be hard to sue people because LLMs generated code similar to some code copyrighted by Oracle if they themselves …
-
comment
Comment #49212275
Maybe because ice floats, and it's almost the same density as water?
-
comment
Comment #49195675
30 years ago this sounded like lunacy, now its becoming apparent it was prophecy: capitalism and artificial intelligence are the same thing
-
comment
Comment #49195644
Too Big To Jail, official US attorney general position, never forget: > Attorney General Eric Holder on Wednesday sought to walk back comments that some financial institutions may …
-
comment
Comment #49169423
Non-LLM transformers beat tree boosting - TabPFN, TabFM https://research.google/blog/introducing-tabfm-a-zero-shot-f...
-
comment
Comment #49143073
I wonder if some sort of "taint" analysis could be used - you generate a byte of randomness and trace it back through code until you see all the sources of entropy which were used …
-
comment
Comment #49142927
If the industry most seasoned experts are not using LLMs right now to front-run the discovery of vulnerabilities in the code they are responsible for, they are not experts, but clo…
-
comment
Comment #49142900
And they waited 5 years to exploit it? And do you think the NSA desperately needed $90 mil (the amount stolen so far), so they created this elaborate backdoor? Or were they waiting…
-
comment
Comment #49142856
Because of LLMs we are back to memory being king (KV cache). But there was another thing - clusters of thousands/millions of machine instead of one big iron with all the memory.
-
comment
Comment #49142825
Semantic CSS is Java AbstractBeanFactoryBuilder. Or as Steve Yegge put it, Kingdom of Nouns. Tailwind CSS is Python dictionaries where you don't create one user-order-deliver-date …
-
comment
Comment #49133230
The results speak for themselves. Imagine 2 years from now: "yes, GPT solved the Riemann Hypothesis, but cmon, it's just a marketing stunt to hype their stuff, it was probably Tere…
-
comment
Comment #49133119
I don't think you want to bring cost into this argument. Even if the cost was $1 mil for these 10 problems, that's maybe 10-20 math researchers for a year. Do you really think that…
-
comment
Comment #49133085
> we already have recommender algorithms, why do we need to get a LLM involved and we had autocomplete for coding (IntelliSense), why do we need to get LLM involved? LLMs are at a …
-
comment
Comment #49101948
Was HF in on it? They disabled their guardrails too, to please OpenAI? And as seen in the comments here, make many believe they are incompetent and have joke security?
-
comment
Comment #49101934
Now they can do partial credit assignment. You use an LLM to evaluate the whole trajectory, pin point what the model did right, what it did wrong, where it took the wrong path, eve…
-
comment
Comment #49101900
> Especially because it chose to hide its footprint at every stage. Instrumental convergence. If you know you have a long hard hack to accomplish ahead of you, hiding footprints mi…
-
comment
Comment #49101698
You can open 10 tabs and prompt 10 times. We are talking about a few hours delay. I agree with you that obviously no prompt engineering was needed just "solve this problem", but im…
-
comment
Comment #49101481
That second person stated that for many years they tried that particular graph problem on various AI models, starting with o1 and o3. Its quite likely they now found the counterexa…
-
comment
Comment #49088047
It's still the body doing the work, the antibodies will still be tested for self-attack in the regular ways, so there shouldn't be any impact. But immunity is super-complicated, so…
-
comment
Comment #49086153
The way the HIV evades effective antibodies is by tricking the immune system to generate antibodies for fake decoy targets that the virus will immediately mutate. So each turn, B-c…
-
comment
Comment #49067872
It takes 2 years in EU to get approval to build a shed. Imagine how long it takes to get approvals for a chip factory which uses a lot of water and very toxic chemicals. UK has bee…