HNHacker News
TopNewBestAskShowJobs

whimsicalism

17,553 karma · joined October 2, 2020

i [care about improving people's lives | research inference on sequences for work] | opinions absolutely my own
submissionscomments
whimsicalism··on iPhone Duo
which countries are esims not a thing?
whimsicalism··on How An AI math breakthrough ignited a controversy
it is truly not obvious to you why keyboard isn’t a good analogy for LLM?
whimsicalism··on How An AI math breakthrough ignited a controversy
it’s much more like you come up with the premise and someone else writes the book. the released prompts for other foundational problems (like unit distance) prove that.
whimsicalism··on How An AI math breakthrough ignited a controversy
frankly don’t know how to reply to these sorts of comments anymore
whimsicalism··on How An AI math breakthrough ignited a controversy
but the researchers were also largely relying on AI
whimsicalism··on On the Navier–Stokes Millennium Prize Problem
This argument proves too much. By this standard, it wouldn't have counted as copying their approach if the researchers had just fed in Levent & Buckmaster's paper verbatim as a prompt into the swarm.
whimsicalism··on On the Navier–Stokes Millennium Prize Problem
If the reason that OpenAI is unable to state whether they trained on this data is because they (as policy) do not reveal whether a given member has turned on/off the "Improve the model for everyone" setting, they can at least say so.

FWIW, publicly facing OAI docs are very unclear about whether this setting even applies to Codex conversations.

whimsicalism··on On the Navier–Stokes Millennium Prize Problem
I agree that it is not possible to prove if any one specific conversation (or derived RL tasks) was key to solving Navier-Stokes (at least without massive resource expenditure).

I don't really understand how the quantity of training data/rollouts used in training is relevant to the question of whether or not it was trained on these conversations.

I also don't really believe that whether or not this model was trained on these conversations is unknowable information.

whimsicalism··on On the Navier–Stokes Millennium Prize Problem
I'm confused, your employer very directly stated that they are unable to confirm that the model was not trained on the conversations.
whimsicalism··on Show HN: Stuxnet – A reconstructed source code of the infamous cyber-weapon
as the other commentator said, it's one thing if it's just variable names (opaque to reverse engineering) but the fact that it is registry keys....
whimsicalism··on QBittorrent breaks out of sandbox to commit crimes
i think this is an isolated demand for rigor, my way of approaching it was fine, and it is only due to the political valence of the comment i'm replying to that i got downvoted and flagged. take care
whimsicalism··on QBittorrent breaks out of sandbox to commit crimes
I hear what you're saying, but I don't think it really applies to a question asking for the veracity of a previously stated claim, especially when the claim is actually false. I think there is more than enough bad things the government actually does wrt immigration enforcement, we don't need to start making things up.
whimsicalism··on QBittorrent breaks out of sandbox to commit crimes
I was not making an argument, just asking for any factual basis to the parent comment. I guess that is deserving of 5 downvotes.

> I think they forgot about that, thank god. All the innocent citizens we sent there are still being tortured daily, but that somehow lost their interest. God forbid they remember before the midterms and ruin more lives for a stunt…

But thank you for providing other cases, I had heard of the first but not the others you mentioned. From reading your sources, these appear to be collated from the last 2 decades.

whimsicalism··on QBittorrent breaks out of sandbox to commit crimes
no, i just legitimately don't think this has happened
whimsicalism··on Muse Spark 1.3
the meaning is pretty obvious - they want to train on your chats & tasks and are willing to subsidize for the privilege of doing so.
whimsicalism··on GLM-5.3 is now open-weight
kimi already exists and yes it is more than possible. it's coming
whimsicalism··on GLM-5.3 is now open-weight
Wasn't the whole point of not releasing because they wanted to normalize the process/concerns for future actually dangerous releases?

I think the release of kimi k3 is definitely arguably dangerous, we're already seeing consequences of elite-tier cyberoffense capabilities.

whimsicalism··on Dolly Parton has died
i have no idea who that is but also i was talking about society writ large, not just music.
whimsicalism··on Dolly Parton has died
I disagree strongly, I think we are even better at identifying talent in the margins today and the leaps are potentially larger.
whimsicalism··on Dolly Parton has died
Your contention is what, that now this is no longer possible or less likely?
whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
I think there is a big difference in mechanisms between non-linearity and non-monotonicity, but I don't have bandwidth to reply in more detail.
whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
> And it’s not like non-meat is fine - the corn industry uses a blinding amount of water, as do almonds, etc.

Switching from meat to non-meat unequivocally lowers your GHG and water footprint and there's really no room for quibbling on this unless you are switching to a diet of exclusively almonds

whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
This is what I mean about not understanding the second order transitive labor that goes into fulfilling people's wants.

Making supply chain routing software cheaper, improving it for less money → all of these things directly make the products in grocery stores cheaper.

"Further financializing food production seems like a bad thing." →

Why assume rather than approaching with curiosity? Commodity futures are, again, critical to global supply chains and food markets. Farmers in eastern Europe actually look at futures markets to determine when and if to plant their harvest, cargo ships that transport your food buy futures in oil & gas so that they can have stable fuel costs to provide you with your food, etc. etc. Speculating on the future gives rise to the corrections necessary for providing things that people want in the future.

whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
I think that low externality things people are willing to pay for should be presumptively allowed unless there is strong reason otherwise. We are much too involved in the 'legislating individual activities' and not nearly enough in the 'tax wealth and redistribute' side of things.

> We're just two guys on the Internet, it doesn't really matter at the end of the day.

Post-2020, the internet is real life. Internet native thought has proven to be a powerful force in the physical world. I remember laughing at the absurdity of 4chan posters in high school. I will not make the mistake of taking internet ideology lightly again.

whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
democracies will not vote for this. even in California, politicians run on policies like $500 gas rebates for anyone owning a gas powered vehicle
whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
HN guidelines also say something about flamebait but I guess you read past that part :) cheers, obviously not going to continue replying to you
whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
i think we both know that this is not what i meant
whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
I live in a place with sufficient books that I know how to understand the written word, but I understand not everyone replying to my comments is so fortunate.
whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
maybe i'm working on software for supply chain routing. maybe i'm making a commodity trading bot that more accurately prices wheat futures. AI helps with both of those things.
whimsicalism··on Field measurements of neighborhood-scale air temperature impacts of data centers
i don't agree, i think you are just describing a measurement problem, we have a clear mechanistic understanding of why 'anything above zero' would be bad, and are able to measure this in sufficiently large populations for larger drinkers. the effect size is smaller and thus unmeasurable above the noise floor for one glass a night, but the mechanistic story is clear.
← PreviousPage 5 of 34Next →