820 karma · joined December 27, 2020
"If you tax the rich, they will leave" is a myth created by the rich so that you won't tax them.
In the middle of the 1900s, the tax rate on the wealthiest Americans was 91%.
Nobody is going to leave over Mamdani's extremely modest proposal.
They're relevant because this was almost certainly written by a PR firm being paid by Musk to resuscitate his 32% approval rating.
However, part of designing and upholding a safety-critical software development process is looking for places to reduce or eliminate the introduction of bugs in the first place.
Strong type systems, for example, eliminate entire classes of errors, so mandating that code is written in X language is a pro-active process decision to reduce the introduction of certain types of bugs.
Restricting the use of AI tools could very much be viewed the same way.
This handwaves away the most important fact w.r.t. stocks and wealth distribution, which is that wealthy people own stocks and poor people don't. A whopping 38% of Americans don't own any stock.
And those equity grants are effectively as good as cash, since they are publicly tradable stock.
I agree, but that number for most people is not the $207k/yr Oxide is paying.
For most people that number is likely north of $500k, if not single digit millions.
That's why these reports tend to suggest corrective actions to the parts of the system that didn't work properly. Even in a perfectly functioning safety culture, an employee can make a mistake and forget to install the bolts. A functioning safety system has safeguards in place to ensure that mistake is found and corrected.
Why are people who prioritized public health during a temporary pandemic "sheep", but the President's supporters who demonstrate exceptionally cult-like behavior, "free thinking individuals"?
You can't have it both ways.
The point that was trying to be made, which I agree with, is that anthropomorphizing a statistical model isn’t actually helpful. It only serves to confuse laypersons into assuming these models are capable of a lot more than they really are.
That’s perfect if you’re a salesperson trying to dump your bad AI startup onto the public with an IPO, but unhelpful for pretty much any other reason, especially true understanding of what’s going on.
Some of the comments I hear almost universally from prior Model 3 owners when they switch to an Ioniq 5 is how much nicer the ride quality is and how nice it is to have buttons on the dash again.
I don't come from poverty, I come from a firmly middle class background. We were a single income household where my dad was a public attorney. Nobody in my immediate or extended family worked in tech. Over the course of my ~15 year career, I've built up a fairly extensive network of former coworkers, many of who I'm sure would try to hire me or get me referrals at their companies if they found out I was on the market. None of this was built through nepotism, as I literally had no connections in tech when I started out.
So, that's the question. The author claims they have had a 20 year career. What happened to all those connections? Do they have a bunch of connections, but no prior coworkers would want to work with them again?
Having not-starving kids is beneficial to literally everyone in the community, not just the kids who use the program.
> When measured statistically.
THANK YOU. It's mildly infuriating how often people forget that one of the things most human experts are good at is knowing when they are looking at something that is likely in distribution vs. out of distribution (and thus, updating their priors).
IP isn't ownership over an idea, it's ownership over a specific artifact.
You cannot copyright the idea of painting or the ideas contained within a specific painting, but you can copyright this specific painting because this specific painting is an artifact.
For example, suppose English had a specific exception such that astronomer is always to be preceded by “a” rather than “an”. The model would learn this simply by observing that contexts describing astronomers are more likely to contain “a” rather than “an” as a next likely character, no?
I suppose you can argue that at the end of the day, it doesn’t matter if I learn an explicit probability distribution for every next word given some context, or whether I learn some encoding of rules. But I certainly feel like the prior is what we’re doing today (and why these models are so huge), rather than learning higher level rule encodings which would allow for significant compression and efficiency gains.
Whoever designed the POTA website… it’s uncharacteristically brilliant for the amateur radio community. There are gazillions of metrics you can track about which parks you’ve hunted and which ones you’ve activated, progress bars for every state, all sorts of awards and “achievements” for various operating times, modes, repeats, etc.
It’s turned portable shortwave operating into gamified crack, except these are real skills that are valuable during an emergency. Having the equipment is one thing, but the regular practice of knowing how to quickly set it up and operate it anywhere is invaluable.
I would fully support a 100% wealth tax for everything over $1B.
Now, of course, the implementation of such a thing would be enormously complicated. There are a million places to hide your wealth, squirreling it away behind shell companies and dark LLCs which you or a designated agent are a controlling member of.
But in concept, I cannot fathom why a human being could possibly say, "You know, $1B really isn't enough for <this>."
Perhaps if Tesla wants to be treated like a regular car company they should get a regular car company CEO?
Funding for sure. :)
But as for perception, the inverse is also true. If I have an perfect planning/prediction system, I can throw the grungiest, worst perception data into it and it will still plan successfully despite tons of uncertainty.
And therein lies the real challenge of robotics: It's fundamentally a systems engineering problem. You will never have perfect perception or a perfect planner. So, can you make a perception system that is good enough that, when coupled with your planning system which is good enough, you are able to solve enough problems with enough 9s to make it successful.
The most commercially successful robots I've seen have had some of the smartest systems engineering behind them, such that entire classes of failures were eliminated by being smarter about what you actually need to do to solve the problem and aggressively avoid solving subproblems that aren't absolutely necessary. Only then do you really have a hope of getting good enough at that focused domain to ship something before the money runs out. :)
Of course not. Enraged, uninformed people "engage", and that sells ads like hotcakes.
I don't know where people get this idea that Zuckerberg had any principles or gave a shit about anyone but himself. He's spineless, and his primary goal in life is has always been acquire as much wealth as possible by whatever means necessary.
There is absolutely no meaningful signal about a system’s safety that can be derived from one person using a system for two weeks.
At best it can only demonstrate that a system is wildly unsafe.
There is a very large chasm of 9s between one person being able to detect an unsafe system in two weeks of use and actually having a truly safe system.
> This time it's different. The naysayers are wrong.
This has been said every time. What is fundamentally different this time?
> LLMs today can already automate many desk jobs.
Which desk jobs existed 2 years ago that don’t exist today because LLMs have automated them away?
Assuming that all technology should imitate nature is a naive engineering principle. The solution should solve the problem within the given constraints.
How would a "function approximation" of Newtonian physics, with billions of parameters, be cheaper to compute?
It seems like this would both be more expensive and less correct than a proper physics simulation.