HN user
Longing for a time before eternal September.
Are you saying that it's slow for training? Sorry, your comment is confusingly worded to me.
I've not had any issues in that regard, but I'm working with LLM's not training diffusion models. Are you following one of the Nvidia provided recipes or inventing something on your own? The last time I looked into it they benchmarked very well, but we both know that doesn't always mean much.
If you bought it for inference you made a mistake. They aren't good at that. Use it to train models and experiment with ML. It's much better at that.
If you just want local inference buy a Mac.
If you bought early on, like I did, the Spark is probably worth double what you payed now. I think I paid $3,000 retail for mine and the last time I looked they were fetching close to $6k on ebay. I'm not sure if that's still the case, but you can buy a very nice Mac with $6k.
If turning other peoples copyrighted work into a model is transformative enough to be protected then so is distilling that model into a different, better, model.
This is very misleading and almost certainly deliberately so.
Texas is the only is State that doesn't participate in the national grid (90% anyhow). Their prices are lower because they dont abide by the regulations that they would be required to if they did. Like weather proofing infrastructure.
This is also why their grid crumbles every time it gets cold, or hot, or... Tuesday.
See the 2021 grid failure for a particularly bad example. In that instance somewhere between 250-700 Texans died to keep their prices low.[0]
This seems to show that even with China taxing industrial electricity to subsidize household electric bills it's still 30% cheaper for industrial electricity in China. [0] Those same taxes make household electricity about half the price of US power.
If I had to wager why, I'd say it's due to embracing solar on massive scales recently. Only a few years ago the US was competitively priced.
[0] https://www.globalpetrolprices.com/compare_countries/USA/Chi...
Its the same root cause. The asker has no "theory of mind".
Often they are neurodiverent. Sometimes they're just overwhelmed with whatever it is they're asking about and dont have time to consider other peoples perspectives.
Its why after decades of having Google in everyone's pocket people still post questions a simple Google search would have solved.
We should all operate under that assumption
We've all had access to Google for decades. People still don't Google things.
If you go around assuming that other people are competent and did the work you would have done you will be disappointed a significant percentage of the time and waste a lot of everyone's time.
You should probably assume your colleague has tried before asking for help,
Spoken like someone who has never worked in a customer support role.
You can never assume competency, from anyone. Even the best amd brightest have bad days.
I too have access to LLM’s.
You should have stated what you already tried then. You didn't, so they did the obvious thing for you... to do you a favor. If you'd tried something before asking them surely you'd have told them that at the start, right?
as evidenced by all the slop dumping
I see more evidence of lazy questions. A good question includes what you've already done to help understand the issue (often refereed to as the proof of work). If you don't do that the other party must assume you are either don't now how, or didn't have time to. If they are kind they do the obvious thing for you. If they are not kind, they just ignore your lazy question.
I mostly ignore the lazy ones these days. Tell me what you tried or I assume you're just trying to get me to do the reading so you don't have to.
You should value the people who take the time to answer your inadequate question at all, rather than bemoaning them for not doing ENOUGH free labor for you when you weren't even willing to type a few lines about what you'd already tried.
Now if you say "I've already Googled it and asked all the AI's, but can't find anything" and they still dump some slop on you then there's a problem.
I suppose that "civil" is relative. Personaly, I wouldn't speak to someone who just tried to help me that way.
Instead I would restate the question to avoid bad answers. If you start by saying what you've already tried, and how it worked, it cuts off most of those "lazy answers."
In my experience most lazy answers are provided as response to lazy questions.
That doesnt excuse you from the obligation to be civil. Especially given that the other party was trying to help.
There are some people I'm not friends with anymore who didn't understand conversational bidding.
To you the "bid" feels like a genuine attempt to connect, that may be a bit indirect.
To some people it feels more like you are being disingenuous or awkward. They instinctively feel that your intent doesnt match your words and it sets off alarms bells.
The fact that you're on the internet judging the other party negatively for having a different communication style is an indicator of which party was least willing to adapt.
Another human being took time from their life to try and help you
It wasn't as much time as you felt entitled to from them, so you responded by minimizing their contribution and being condescending. That upset them.
The problem here was not the other person.
Just for those who aren't aware, most of the recent push for these laws has been bankrolled by Meta. They wish to avoid legal responsibility for their attacks on democracy and human health by convincing governments that people don't need any right to anonymous speech, and thusly free speech, as much as Meta needs to not have to pay moderators.
All smart people are curious about things outside their realms of expertise. Curiosity is one of the defining attributes of intelligence.
The desire to discuss those things is also reasonable. Though I would agree that those without credentials should probably abstain from expressing any potentially harmful opinions when healthcare is involved, the freedom to ask questions or bring up interesting digressions is fundamental.
I'm gonna get a few people together and all run for city council so we can seize profitable domain names for ourselves.
"Sorry Meta, but BFE, Nebraska outlawed Farmville and now some guy named Bob owns facebook.com."
This is the worst they will ever be.
In 1956 a 5mb hard drive shipped on a large truck and took a team of men to unload. It consumed huge amounts of power, and cost about $3,200/month to run. In today's dollars that would be about $160,000 per month.
Aren't you glad we didnt just give up because it was kind of expensive?
I've always been fond of calling it telepathy. That's really what it is. Further, it can even let us see into the minds of the long dead. Humanities superpower is that it is the only race which can preserve knowledge beyond a single lifespan.
With LLM's we can also now preserve actual thought patterns and skills beyond a single lifespan. I can't wait to see what the future brings.
I think the words are just a medium for thoughts. AI only produces the former, not the latter.
I think this is exactly what I was trying to say above when I talked about perspective Reading is a form of telepathy, and there's just nothing INSIDE of most AI writing today. When that changes, I'll take it more seriously. I do suspect that it will very soon though.
As you say, AI's are already coming up with novel math and I've seen some glimmers of purely mechanical brilliance in the writing already.
I once read Ben Franklin's autobiography. I have yet to experience any great blossoming of our relationship.
More seriously, the novel is a unidirectional medium of communication. Not a relationship. That said, it is meant to convey a perspective. I don't care if it's a human, a robot, or a little green man's perspective. It just has to be an interesting, useful, or enlightening perspective that I didn't already share.
Right now LLM's don't really have unique perspectives, that may change. So I withhold judgement. As of today, I wouldn't read an LLM's novel. In 10 years, who knows?
Anthropic and Google already charge extra to use your own harness. That's 100% of the reason I'm using OpenAI.
If they go down that path I'll just go back to my old buddy Claude, or maybe buy a second Spark and keep it local.
Your economic value and your value as human being are two different things.
Yeah, you gotta eat, but some of us have been slaves for so long that we've started to believe that our self worth is defined by what we will fetch at auction. It is not.
Read more about the Protestant work ethic, it's history, and what life was like before men began to assume it was the one true view of life. Not very long ago nobody had clocks, and the idea that work was an ethic was a fringe, almost cult-like belief system. Now most humans routinely take stimulants every morning to be better servants, and mistake their job for their purpose on earth. That's just nonsense our culture invented though.
If the master wants you to do something, force him get out the whip. Don't let him live in your head rent free by adopting his crazy belief system.
YOU decide what your value is. Don't abdicate the decision to someone else.
If there was a Karpathy style universal solver, it wouldn't be very smart unless we scaled it up.
I think that the scaled up version is actually still really valuable.
Imagine being able to just add more compute as needed for any given problem until it's solved by just adding more copies of a single universal layer, without more training. Or being able to burn the individual core into silicon and just loop it as needed.
I tried to build exactly that in my personal lab once, but hit a wall made of my own incompetence and budget.
The idea was to find the parts of the manifold that did generic reasoning and then scale as needed by repeating them. It worked within individual layers (I could make the model score higher on benchmarks by repeating the reasoning extracts within individual layers), but i could never get the interfaces between layers to work again after I'd done that. I suppose it needed traing to "heal" the interface again after my brain surgery, but I didnt have the compute to manage it and moved on to the next project
I'm sure that someone who actually gets paid to do these things will figure out some version of it eventually though, because I know it can be done.
I agree, but maybe for different reasons. I think Karpathy is right. We need models that reason, not models that memorize.
Karpathy calls it a "Cognitive Core", and it's essentially a small model that learns to reason and look up the data it needs as opposed to a giant model that memorizes all the data in the world and tries to process large chunks of it all at once with every thought. I think it will be based on the thing that grokking, the lottery ticket hypothesis, and the universal weight subpspace hypothesis all point to.
Eventually someone will figure out how to build it and the entire economy that we've now built on top of the wacky idea that nothing can possibly ever get more efficient will collapse overnight.
Sometimes I wonder how much Nvidia would pay someone not to release a thing like that, and then I wonder if that's already happened.
Most browsers don't render punycode .coms very well the way we'd hope for these days. Homoglyph attacks ruined the fun.
Every generation has felt some version of this.
"Keyboards are soulless. Handwriting is personal – as unique as fingerprints." - Joyce Carol Oates on typewriters
"This discovery of yours will create forgetfulness in the learners' souls, because they will not use their memories; they will trust to the external written characters and not remember of themselves. The specific which you have discovered is an aid not to memory, but to reminiscence, and you give your disciples not truth, but only the semblance of truth." - Socrates on writing
This is a great question. Thanks for asking it. It really got people talking.
My $0.02: Your value as a person never had anything to do with your value to the economy. It's time to relearn that fact.
Check out some of Tom Hodgkins books for more. Or maybe anxietyculture.com.
It is possible that I do. Fifty percent of people are of below average intelligence.
People have different beliefs about whether they will personally benefit from supporting some political cause.
The key word there is "belief". They are often wrong.
Your linked blog post is backwards and inconsistent with itself. You have two primary arguments: Irrational and Immoral. You argue that voting is irrational because its unlikely to have any impact, and that voting for your own interest is immoral.
A) The statements are mutually exclusive. An act that has no impact on others can not be immoral.
B) It assumes that what is best for the individual is worse for the group. Life is not a zero sum game. That's the Conservative's delusion. Economic and political transactions do not always have a "loser" and a "winner". In fact, it's relatively rare that they do if you think more than zero steps into the future.
C) The only version of this that actually works is the opposite.
C1) It is irrational not to use whatever influence you have to effect you environment for the better, even if the expected value is low because the opportunity cost of inaction may be disasterous. It's similar to your odds of dying by meteor strike. The probablity is higher than you expect because the death toll would be enormous if it did happen. Outlying events with large impacts skew the numbers.
C2) It is immoral to vote against your own interests, because what is best for the group is also what is best for members of the group. Any other belief is just an incorrect belief based on imperfect knowledge. Again, your argument makes sense at step zero, but not at step 'n'. If what you're voting for seems bad for some members of the group, but good for you, it just means you have imperfect knowledge of what's actually good for you in the long term.
Because a percentage of every dollar you spend on it will go towards pushing political opinions that run contrary to your own best interests?