Obviously I think that AI generated undressing pictures of people, especially minors, is bad and there should be safeguards against that. But how is it different from other tools like doing it manually with photoshop? Also it has been shown that many other tools like ChatGPT and Gemini/Nanobanana can also do it with sufficiently creative prompting.
I also did scroll through the public grok feed and the AI generated bikini pics were mostly Onlyfans creators requesting their own fans to generate these pictures (or sometimes generating them themselves).
You know the answer to this but I'll just say it: Its different in that it requires no skill and can be done by anyone instantaneously at scale.
You know this but somehow are rationalizing this game changing fact away.
Yes, people can draw and photoshop things. But it takes time, skill, dedication, etc. This time cost is load bearing in the way society needs to deal with the tools it has for the same reason at the extreme that kitchen knives have different regulations than nuclear weapons.
It is also trivially easy for grok to censor this usage for the vast majority of offenders by using the same LLM technology they already have to classify content created by their own tools. Yes, it could get jailbroken but that requires skill, time, dedication, etc; And it can be rapidly patched, greatly mitigating the scale of abuse.
> But how is it different from other tools like doing it manually with photoshop?
The scale of effect and barrier to entry. Both are orders of magnitude easier and faster. It would take hours of patience and work to mostly create one convincing fake using photoshop, once you had spent the time and money to learn the tool and acquire it. This creates a natural large moat to the creation process. With Groom it takes a minute at most with no effort or energy needed.
And then there is the ease of distribution to a wide audience, X/Groom handles that for you by automatically giving you an audience of millions.
It’s like with guns. Why prevent selling weapons to violent offenders when they could just build their own guns from high quality steel, a precision drill, and a good CNC machine? Scale and barrier to entry are real blockers for a problem to mostly solve itself. And sometimes a 99% solution is better than no solution.
This thing with guns was a legitimate argument for banning or regulating 3D printers a few years ago though and I'm glad that we didn't end up with restrictions on that front. With affordable desktop CNC machines capable of making metal parts coming soon, I hope those won't be subject to too many restrictions also.
I am for moderation and strong penalties for users that use it in that manner. Anyone who uses grok to generate an undressing image of someone without their consent within 5 seconds should probably go to jail or whatever the penalty is for someone spending 5 hours to create revenge porn with photoshop.
But I'm not sure if the tool itself should be banned, as some people seem to be suggesting. There are content creators on the platform that do use NSFW image generation capabilities in a consensual and legitimate fashion.
Grom is much much less censored on purpose. I work in image editing and outside of very few people, hardly anyone uses Grok for professional work. Nano Banana Pro is used for the most part.
But for NSFW work it dominates. It’s clearly deliberate.
If you generate CSAM, whether using LLMs, photoshop, or any other tool, you are breaking the law. This would apply if you could somehow run Grok locally.
When you use a service like Grok now, the service is the one using the tool (Grok model) to generate it, and thus the service is producing CSAM. This would also apply if you paid someone to use Photoshop to produce CSAM: they would be breaking the law in doing so.
This is setting aside the issue of twitter actually distributing the CSAM.
The Photoshop equivalent would be "an Adobe artist does the photoshop for you and then somehow emails it directly to your target and everyone who follows them."
Exactly. And the fact that companies do it with impunity is another hint that we're living in late stage capitalism.
If an individual invented a tool that can generate such pictures, he'd be arrested immediately. A company does it, it's just a woopsie. And most people don't find this strange.
I think intent probably matters and that this gets into the "you know it when you see it" definition realm where we debate the balance between freedom of speech and security of person. ie. just how easy Photoshop, a VCR, a DVD burner app, etc. makes it for you to crime and how much are they handholding you towards criming?
I think this is an important question to ask despite the subject matter because the subject matter makes it easy for authorities to scream, "think of the children you degenerate!" while they take away your freedoms.
I think Musk is happy to pander to and profit from degeneracy, especially by screaming, "it's freedom of speech!" I would bet the money in my pocket that his intent is that he knows this stuff makes him more money than if he censored it. But he will of course pretend it's about 1A freedoms.
Friction/barrier to entry is the biggest difference. People generally didn't do things like that before due to a combination of it being a colossal waste of time and most not having the requisite skills (or will and patience to acquire said skills). When all it takes is @mentioning a bot, that friction is eliminated.
The obvious problem is that Grok is also distributing the illegal images.
This could be easily fixed by making the generated images sent through private Grok DMs or something, but that would harm the bottom line. Maybe they will do that eventually once they have milked enough subscriptions from the "advertising".
How is having cameras on every street corner that identify you based on your face and height and weight and gait and the clothes you're wearing and anything you're carrying, or the car you're driving by its license plate and make and model and color and tires/rims and any visible damage, accessories, etcetera, and taking all these data points and loading them into a database that cross-correlates them with your credit bureau data and bank records and purchase history and social media and other online activity and literally every single other scrap of available data everywhere, and builds a map of everything about you and everywhere you ever go and everything you do and have ever done, makes it trivially queryable by any law enforcement officer in the country with or without a valid reason, retains it all in perpetuity, and does all this for every single person in the country without consent or a warrant issued by a judge, different from a police department assigning an officer to tail you if you are suspected of being involved in a crime?
We are going to be in some serious fucking trouble if we can't tackle these issues of scale implied by modern information technology without resorting to disingenuous (or simply naive) appeals to these absurd equivalences as justification for each new insane escalation.
> But how is it different from other tools like doing it manually with photoshop?
Last I checked Photoshop doesn't have a "undress this person" button? "A person could do bad thing at a very low rate, so what's wrong with automating it so that bad things can be done millions of times faster?" Like seriously? Is that a real question?
But also I don't get what your argument is, anyway. A person doing it manually still typically runs into CSAM or revenge porn laws or other similar harassment issues. All of which should be leveraged directly at these AI tools, particularly those that lack even an attempt at safeguards.
> why Singapore got to be such an important global hub.
Without the the location, of course Singapore wouldn't have been able to be so important. But the location isn't everything --- Singapore manages to outperform Port Klang and Tanjung Pelepas despite the similar geographic advantages of the Malaysian ports due to much better execution.
Given that fully driverless Model Ys and Cybercabs have been spotted going around Austin, I find that the "they are not on a path to becoming a real service" is a little too strongly worded.
* Elon has been making wildly exaggerated and over-optimistic claims for a decade and continues to do so
* Tesla has recently made huge strides in capability and has a clear path to full autonomy
And to be fair, many other car companies also promised self driving cars, e.g. Audi in 2014 promising driverless cars by 2016 [1]. It's just that Tesla is still executing on the promise whereas many other carmakers have fizzled out on their ambitions. As the Rodney Brooks article itself mentions,
> As a reminder of how strong the hype was and the certainty of promises that it was just around the corner here is a snapshot of a whole bunch of predictions by major executives from 2017.
I don't understand how these cars keep getting stalled for half an hour or something. Surely there must be a team of teleoperators ready to jump in at any time?
"the infra" is cell phone data coverage though. Which was probably congested by all the city residents using theirs instead of their wifi which was down. Would be fascinating to see just how much Internet traffic flows changed during the outage.
EREV is different from diesel-electric in that the EREV has a large battery whereas the diesel-electric locomotive does not. But the "ICE engine drives a generator which drives a motor" philosophy is similar in spirit.
Yes true; good point. I think this is changing (e.g. regen braking for aux. power on passenger trains maybe eventually capacitors for traction drives in the future), but currently and ~almost all the time, this is correct and a good point.
I feel like the home depot website is fine. It's a lot better than most other shops, I've had a good experience finding the aisle and location of items, and it's generally accurate with the amount in stock at each location. If you didn't enable precise location or have bad cell signal then that is hardly the fault of the website.
I will not argue with the stock part. When the search _does_ finish, stock info is usually correct IME.
What grinds my gears is the speed of this search, regardless of the phone reception. Even on the desktop it feels like they have a bunch of interns running a sneakernet. Or the website is laden with pointless javascript that slows everything down before the search is actually performed.
I go to the same Home Depot every time. (Well I don't if I can help it, but that's beside the point). There is no reason they cannot store the preferred store in the localStorage or cookies or wherever else. Other stores have figured this out.
Not CostCo though! I open their page and immediately 'Can Costco.ca use your location?" I say yes and then it asks me what province I'm in. I tell it, and then it defaults me to a store 30 minutes' drive from here and not the one five minutes away. Every. Time.
On the contrary, you can also search for items with easily the least useful search I've ever seen.
Pizza pockets? Okay, anything that mentions pizza, or pockets. So frozen pizzas, in-store-made pizzas, pizza flavored pringles, pants with pockets, dresses with pockets, frozen items that are similar to pizzas but aren't pizzas, frozen items which aren't similar to pizzas, granola bars I guess, basketballs, and so on.
It seems as though their search just takes the search terms, matches them against every item in their database in order of relevance, and then just shows you everything regardless of how relevant it is. 0.00121 out of 100? Well, it's still technically relevant! Let's show it just in case!
Their internal setup was also an absolute mess as of 4 years ago. A horrific hybrid of extremely legacy systems and new systems created around COVID which are both nicer and also deeply lacking in features we needed as floor workers.
I understand that upgrading and migrating to new systems takes time but this process never seemed like it involved anyone on the ground.
This is definitely true and makes the experience shittier than it otherwise would be, but even with a great signal/connection it frequently loads so slowly that I've long run out of patience.
I have gotten in the habit of looking up what isle and bay the thing I need is before I get there, and then I screenshot it because too many times the page has needed to reload and start over
I bought a water heater that had a large (1k!) instant rebate that you had to scan, sign up on website and show the emailed coupon to the person during cashing out. Took me 25 minutes wandering around the store to get enough reception to actually do this process. Made me chuckle, thinking how having it online only but before point of sale in the store was such a terrible, terrible idea.
Nah, I use both the website and their shitty web wrapper app on a regular basis and it's been a dumpster fire for at least the last 2-3 years. 3-5 years ago when they first rebuilt everything it was much more pleasant but at this point it's clear no one is maintaining it and have just let it bloat and rot
Rivians have been spotted with giant Velodyne VLS-128 "Alpha Puck"s since several years ago [1]. But from last I checked, Rivian's ADAS is still struggling with ping-ponging in lanes on curved stretches, and it only works on a small set of pre-mapped highways. Highly doubtful that "universal hands free" is coming.
The ping-ponging is certainly a Gen1 problem. (My Gen1 does this.) Gen1 was essentially an off-the-shelf Mobileye unit, and the performance was, as expected, not good.
Gen2 autonomy stack is completely unrelated to Gen1, and from what I hear is a completely different level of reliability.
(also - this presentation covered yet another, unrelated, gen3 autonomy stack, which shares none of the hardware or models with the existing gen2 stack, either.)
The big lidars are for ground truth collection. They get used in projects ranging from autonomous development all the way down to budget adaptive cruise control or parking sensor benchmarking.
I also did scroll through the public grok feed and the AI generated bikini pics were mostly Onlyfans creators requesting their own fans to generate these pictures (or sometimes generating them themselves).
reply