pj_mukh

The legality of Copyright material being present in the latent space of a commercialized foundation model would definitely be an interesting case!

Though most of these cases aren't litigating this. They are charging that the ability to recreate copyrightable material is the fault of the tool and not the tool user[1].

[1]: https://www.darrow.ai/resources/ai-copyright-law

"express and sole purpose of the tool"

Categorically untrue. 100% of all my usage of image generation is for artistic purposes that have nothing to do with the reproduction of or even the mimicry of existing art.

This is a gaping hole in the legal case if the lawyers need this to be true.

“Photoshop doesn't recommend”

Neither does OpenAI? What about an empty chat prompt “encourages” copyright abuse?

I don’t understand, generating (and then monetizing) things that are copyright violations is already illegal.

What we’re taking about is suing the tool that could potentially be used to generate copyright violations. Which seems silly?

Steelmanning your argument a bit more: OpenAI is selling the fact that you could potentially generate copyrightable material as a subscription, because the artwork is somewhere in the compressed data?

Abstracting the underlying tech a bit: If we imagined a printer that could faithfully represent whatever you told it to and you told it to reproduce Da Vinci’s that you later sold as originals while others told it to reproduce original forms of art (successfully), who’s at fault here, you or the printer?

Seems obvious to me it’s the user of the printer, not the printer itself, even if the printermaker is making money.

Sincere question: Do we need new laws for this?

If I use photoshop to create a straight reproduction of an artwork by someone else and monetize it, copyright law handles this.

If I use AI to do the same, same deal? At no point did someone suggest suing Adobe for building photoshop for "enabling" people to recreate art and sell it.

So what changed? The ease of use?

"Most people see a number $6 Trillion and something breaks inside them. What they don't realize is maintaining infrastructure is expensive."

It breaks all our brains too, because that $6 Trillion has very little to do with maintaining infrastructure. The bulk of it is just direct payouts (Social Security, Medicare and defense contracts).

Pretty frustrating that this big number means the government is politically forced to do drastic austerity for things like keeping planes flying safely.

Also, makes DOGE starting at USAID (<1% of budget) look especially incompetent.

I really think the new AI panic is not as important as the original social media/algorithmic feed addiction panic. In fact some of the core problems with AI is that it’ll be used to generate slop for the algorithmic feed.

“The feed” is the tech world’s original sin, always will be. I wish there was someone around Mark Zuckerberg that would tell him that he needs to lead a movement to close the Pandora’s Box he’s opened. Be like Oppenheimer, Mark. You’ve become the destroyer of cognitive power worldwide, you should lead the movement to end it.

Everything else we’re trying is lipstick on a pig.

Welcome to the defining paradox of the 21st century:

https://en.wikipedia.org/wiki/Moravec%27s_paradox

So if I’m in a protected bike lane with a row of parked cars to my left wouldn’t it be flagging every parked car as a potential hazard?

Orrrr, you can just enforce fare evasion (passively) and mostly solve this problem and solve the other societal problems separately. It’s been tried and it works [1]. I really don’t like everthing bagelling all problems.

[1] https://www.sfchronicle.com/bayarea/article/bart-fare-gates-...

Is squeezing into the 7th seat in the trunk of an Uber-XL SUV worth the faux-instantaneous nature of this product?

Get the Transit app folks [1], great GPS tracking of the right bus for you, advocate for an efficient well-funded bus/train service in your city and a municipal DOT that doesn't have to host 5 community meetings for every small change in routes.

[1]: https://transitapp.com

P.S: I'm an Uber One member, it certainly has its place in a car-less life. But this ain't it Chief.

"The staff has grown 55 percent since Dr. Hinton’s forecast of doom, to more than 400 radiologists."

Wonder what other forecasts of doom he is wrong about :|.

Interesting product! One problem: I want to get a handle on the quality of online conversations you flag. Do you know that on the free tier, the top 5 you flag are the best of your filter?

Clines agent work is better than Cursors own?

They are doing that in East Bay [1] on BART owned land. They had been blocked so far by the most virulent NIMBY trash leading to some of the most scathing coverage [2]. Most have been unblocked and are under construction. The one still being blocked is Ashby BART because of (say it with me folks), a "HIsToriC flEa MARkeT" (that mostly sells stolen goods)[3].

[1]: https://www.bart.gov/about/business/tod

[2]: https://www.nytimes.com/2023/05/30/magazine/housing-berkeley...

[3]: https://www.berkeleyside.org/2024/09/17/berkeley-ashby-bart-...

Otherwise known as the Worst Transit Project in the US [1]

[1]: https://www.youtube.com/watch?v=LZrrtF8Iy8k

So cool! I wonder how the Lidar and ARCore poses were cross-calibrated?

Just to avoid this, I would just use a LiDAR equipped iPhone Pro, with industrial grade cross-calibration and still have all the visualization fun.

Also a lot less robust depending on baseline.

AFAICT the technologies been around for a while (esp assistance). It's the regulation and the minefield of a roll-out plan that the FAA would rather not take on. Nobody gets fired for doing nothing and continuing to play whack-a-mole with human frailties.

I mean maybe instead of patholigizing to that level we maybe need to accept that there is a temporal normal distribution to human attention spans and design our systems around it.

It feels like semi-autonomous ATC and flight controls were possible as of 5 years ago. Has FAA even started writing initial reports on this?

I don't know. Even with these tools, I don't want to be doing this work.

I'd still hire an entry level graphic designer. I would just expect them to use these tools and 2x-5x their output. That's the only changing I'm sensing.

But by then 25M would have access to the main trunk of the train generating massive revenues. That's the point.

Holding up 25M people to try and include 1M more while bloating costs and showing no progress is the essence of how you kill a project.

"they're going to build the train between Bakersfield and Los Angeles through the Tehachapi Mountains."

OMG, just join the I-5 after Bakersfield. It's right there. Why are we barreling through the mountains?

This is the most California article ever written. The author admits that if the line was run along the I-5 it probably wouldn't need all the costly grade-separation and state property issues it's now facing. It may have even finished with its original funding! But instead of immediately being able to serve the upwards of 18M people of Greater Los Angeles and 7.5M people of the Bay Area, the 1M people of the Central Valley would have to wait for branches to be built.

SO INSTEAD we took the more circuitous route through Central Valley so that the 1M people feel immediately included and NO ONE is getting a high speed rail.

Sir! ChatGPT couldn't come up with a more California scented boondoggle.

I think it's actually the opposite. The latent demand is only present in LA and SF. Fresno and Bakersfield are not car-less cities. Taking a train between them makes no sense (you're gonna have to rent a car on either end).

While connecting the LA Metro and SF BART with a high speed rail line makes the most sense. Regardless of the construction inefficiencies this is really the original sin beating at the heart of this project.

Super cool, and will definitely check it out.

But as a measure for what you can achieve with a course like this: does anyone know what the max tok/s vs iPhone model plot look like, and how does MLX change that plot?

California tried to do this, the bill got watered down in committee [1]. It's probably true that purely GPS-based speed-limiting is not good enough. Imagine being on a 75mph highway with a 25mph service road right next to it and the GPS not knowing the difference.

Still, interesting idea that could have legs when the technology got better.

[1]: https://www.npr.org/2024/09/05/nx-s1-5099205/california-tech...

You made the right call. I would never invest in this company, and neither would YCombinator [1]

[1] https://www.ycombinator.com/library/5x-how-to-split-equity-a...

The last 5 years have been a supersized lesson in Moravec’s paradox [1]. More people need to hear about this: the tech world doesn’t have some kind of conspiracy against the creative class, it literally is the lowest hanging fruit as predicted 30 years ago.

[1] https://en.m.wikipedia.org/wiki/Moravec%27s_paradox

Along these lines wouldn’t it be better to require ships to add some forward facing underwater radar to watch out for whale strikes? Or am I missing something here?