The thing is when gasoline is used to power a car about 80% of the energy is loss to heat and only about 20% is used to actually move the car on average. It’s worse when you are doing acceleration and better when cruising at a steady speed on a level surface.
Google web services take advantage of an API that only Google knows about.
Completely unsurprising. Google should have been given the anti-trust treatment long ago. There’s not a saving us because the ones to save us are completely complicit. And people who write independent browsers will be smacked back down by having places like YouTube throttle them.
it physically lives in your RAM for the duration of the stream.
It physically lives encrypted in your RAM and only temporarily. Remember TPM exists.
Maybe that’s faulty, as I haven’t tried it myself
Nah perfectly fine take. Each their own I say. I would absolutely say that where it is, not bothering with it is completely fine. You aren’t missing all that much really. At the end of the day it might have saved me ten-fifteen minutes here and there. Nothing that’s a tectonic shift in productivity.
deleted by creator
In the tech industry (likely every industry but I wouldn’t know) we could make a 200 level course in college that covers tech that’s been over hyped. A few choice hits like:
Now some of those went on to become useful concepts, but all hardly lived up to the hype of transforming the industry forever. There’s just no shortage of people who lack any kind of set of morals that will, without any knowledge in the domain, jump on some train and hype it to get some quick cash before the thing derails in a fit of coming to terms with reality.
I mean, at least it’s been this way since I’ve been in the industry.
I had my fun with Copilot before I decided that it was making me stupider - it’s impressive, but not actually suitable for anything more than churning out boilerplate.
This. Many of these tools are good at incredibly basic boilerplate that’s just a hint outside of say a wizard. But to hear some of these AI grifters talk, this stuff is going to render programmers obsolete.
There’s a reality to these tools. That reality is they’re helpful at times, but they are hardly transformative at the levels the grifters go on about.
Tech vendors have also been falling over each other to tell the world how they are including GenAI in their offerings as the leading AI companies attract feverish attention from investors.
Because you can’t hype it up for investors if you call it what it actually is. Fancy auto complete. And don’t get me wrong, I love me some of the tools out there. But this stuff is being absolutely way over hyped.
It’s good to go into this stuff with realistic views. Will it do all your work? Absolutely not. But what it will do is do a lot of heavy lifting for you so that you can get more things that require your specific attention done.
The level of “sky is falling and we’re all going to be enslaved by AI” is literal bullshit to sell more stocks and create a bubble that will absolutely pop.
AI-generated artwork is detrimental to the creative industry and should be discouraged
Man you wouldn’t guess how airbrush artist felt when Photoshop came around.
I agree unregulated AI is problematic. At the same time, I’m cynical on what the actual measures would look like.
OMG, Thank you, this is the correct take.
China is one of the leading producers of solar panels at exceptionally cheap prices. China and the Philippines are currently in a bit of a sour mood with each other. The US could step in with cheap solar panels if they were researching that, but alas, who do you think is going to be providing them with cheap LNG?
So while the story indicates that solar is at it’s cheapest it’s ever been. Not for the Philippines. The Philippines is a clear cut example of how the United State’s continued resistance to a domestic solar panel production industry is literally hurting their long term prospects. There’s a lot of nations out there that want panels and right now, China is the only serious game.
When folks talk bad about GIMP.
Say no to SaaS as much as you can
I love GIMP and I will die on that hill (yes, fully aware of the things it lacks, thank you). But for those who use Adobe products, from what I can tell, the answer is that they have no choice in the matter. Adobe is just that ubiquitous in that industry that you either use it or you don’t work in that profession.
With Adobe dipping into AI stuff, I have an underlying fear they’re going to become as ubiquitous in that domain as well, that people trying to compete with them just won’t be able to. And then we will have the same problem in AI with Adobe as we have with Digital Image Editors and Adobe.
Ah. No problem. So the notion behind the “big guys are the ones that stand to profit from AI regulation” is that regulation curtails activity in a general sense. However, many of the offices that create regulation defer to industry experts for guidance on regulatory processes, or have former industry experts appointed onto regulatory committees. (good example of the later is Ajit Pai and his removal of net neutrality).
AI regulation at the Federal level has mostly circled “trusted” AI generation, as you mentioned:
But what it is doing is making it infinitely easier to spread enormous amounts of completely unidentifiable misinformation, due to being added with indistinguishable text to speech and video generation
And the talk has been to add checks along the way by the industry itself (much like how the music industry does policing itself or how airline industry has mostly policed itself). So this would leave people like Adobe and Disney to largely dictate what are “trusted” platforms for AI generation. Platforms that they will ensure that via content moderation and software control, that only “trusted” AI makes it out into the wild.
Regulation can then take the shape of social media being required to enforce regulation on AI posts, source distributors like github being required to enforce distribution prohibitions, and so on.
This removes the tools for any AI out of the hands of the public and places them all in the hands of Adobe, Disney, Universal, and so on. And thus, if you wanted to use AI you must use one of their tools, which may in turn have within the TOS that you can not use their product to compete with their product. Basically establishing a monopoly.
This happens a lot in regulatory processes which is why things like the RIAA, the MPAA, Boeing, and so on are so massive and seemingly unbreakable. They aren’t enshrined in law, but regulatory processes create a de facto monopoly that becomes difficult to enter because of fear of competition.
The big guys, being the industry leaders, in a regulatory hearing would be the first to get a crack at writing the rules that the regulatory body would debate on. In addition to the expert phase, regulatory process also includes a public comment, this would allow the public to address concerns about the expert submitted recommendation. But as demonstrated back in the public comment of the debate to remove rules regulating ISPs for net neutrality, the FCC decided that the comments were “fake” and only heard a small “selected” percentage of them.
side note: in a regulatory hearing, every public comment accepted must be debated and rationale on the conclusion of the argument submitted to the record. This is why Ajit Pai suspended comments on NN because they didn’t want to enter justification that can be brought up in a court case to the record.
The barrier is no longer “you need to be an artist”. It’s “you need to have an internet connection”
And yeah, that might be worth locking AI out of the hands of the public forever. But it doesn’t stop the argument of “AI taking jobs”. It just means that small startups will never be able to create jobs with AI. So if the debate is “AI shouldn’t take our jobs, let’s regulate it”, that will only make it worse in the end (sort of how AWS has mostly dominated the Internet services and how everyone started noticing that as not being incredibly ideal around 2019-2021 when Twitter started kicking people off their service and people wanting to build the next Twitter were limited to what Amazon would and would not accept).
So that’s the argument. And there’s pros and cons to each. But we have to be pretty careful about which way to go, because once we go a direction, it’s pretty difficult to change directions because corporations are incredibly good at adapting. I distinctly remember streaming services being the “breath of fresh air from cable” all the way up till it wasn’t. And now with hard media becoming harder to purchase (it’s not impossible mind you) we’ve sort of entrenched streaming. Case in point, I love Pokémon Concierge, it is not available for purchase as a DVD or whatever (at least not a non-bootleg version), so if I ever want to watch it again I need Netflix.
And do note, I’m not saying we shouldn’t have regulation on AI, what I am saying is that there’s a lot for consideration with AI regulation. And the public needs to have some unified ideas about it for the regulatory body’s public comment section to ensure small businesses that want to use AI can still be allowed. Otherwise the expert phase will dominate and AI will be gone from the public’s hands for quite some time. We’re just now getting around to reversing the removal of net neutrality that started back in 2017. But companies have used that 2017 to today to form business alliances (Disney + Hulu Verizon deal as an example) that’ll be hard to compete with for some time.
Honest question. How does their service “replace” an open source LLM? If I’ve got locallama on my machine, how does using their service replace my local install?
From Cara:
We do not agree with generative AI tools in their current unethical form, and we won’t host AI-generated portfolios unless the rampant ethical and data privacy issues around datasets are resolved via regulation
Okay I wanted to talk real quick about this aspect. Lot’s of folks want AI to require things only held in copyright. And fine, let’s just run with that for sake of brevity. Disney owns everything. If you stick AI to only models which the person holds copyright, only Disney will generate AI for the near future.
I’m just going to tell you. The biggest players out there are the one who stand to profit the most from regulation of AI. And likely, they’ll be the one’s tasked by Congress to write drafts of the regulation.
In the event that legislation is passed to clearly protect artists, we believe that AI-generated content should always be clearly labeled, because the public should always be able to search for human-made art and media easily
And the thing is, is Photoshop even “human-made art”? I mean that was the debate back in the 90s, when a ton of airbrush artist lost their jobs. And a large amount of Photoshop that was done, was so bad back then we had the whole Ralph Lauren, Filippa Hamilton thing go down.
So I don’t disagree with safe from AI places. But the justification of Cara’s existence, is literally every argument that was leveled at Photoshop back in the 90s by airbrush artist who were looking to protect their jobs and failed because they focused way too heavily on being anti-Photoshop that the times changed without them. When they could have started learning Photoshop and kept having a job.
I think AI presents a unique tool for artist to use to become more creative than they have ever been. But I think that some of them are too caught up in how CEOs will eventually use that tool as justification to fire them. And there’s a lot of propensity to blame AI when it’s the CEO’s writing the pink slips, just like the airbrush artists blamed Photoshop, when it was newspapers, the magazines, and so on that were writing the pink slips.
I just feel like a lot of people are about to yet again get caught with their pants down on this. And it’s easy to diss on AI right now, because it’s so early. Just like bad Photoshop back in the 90s led to the funny Snickers ad.
Like I get that people building models from other people’s stuff is bad. No argument there. But, open models, things built from a community of their own images, are things too but that’s all based on the community and people who decide to be in a collaborative effort to provide a community model. And I think folks are getting so hung up on being anti-AI, that it’s going to hurt their long term prospects, just like the airbrush folks who started picking up Photoshop way too late.
There’s not a stopping Disney and the media companies from using AI, they’re going to, and if you enjoy getting a paycheck, having some skill in the thing they use is going to be required. But for regular people to provide a competitor, to fight on equal footing, the everyday person needs access to free tools. Imagine if we had no GIMP, no Kitra, no Inkscape. Imagine if it was just Adobe and nothing else and that was enforced by regulation because only Adobe could be “trusted”.
This court is absolutely raring to go on major questions. In the past, when Congress left things wide open, the High Court usually gave deference to the agency to handle the details. Examples are things like:
Congress doesn’t list in massive detail every single possible permutation that’s possible in law. That would create thousand page laws. But as EPA vs WV has shown us, the Supreme Court wants incredible detail. So we get the over 300 pages of new law that indicate six gases, fifteen different levels of municipality, and over ten thousand different industries plus all the various ways those three things interact with each other, to address what was “missing” from the original grant of authority for the EPA.
And the thing is, Republicans will bemoan these large tomes of text, saying “how can we know what’s in it?” That’s them breaks. If the Supreme Court say “a government agency can not do XYZ because it doesn’t say XYZ in the law” then that means we have to be very detailed about what’s in the law. That’s how we get thousands of pages per law. That’s kind of the reason why prior Courts didn’t harp on this stuff. The President changes every four to eight years, regulation can change at that rate too. Law change very infrequently. So that whole EPA vs WV result, CO₂ regulation was something that basically bounced every time we swapped parties, NOW it’s in law and it’s going to be there for decades.
The ISPs are getting ready to shoot themselves in the foot here. Because if NN is enshrined in law, NN is here to stay. As long as it’s a regulatory process, it can change President to President. But push come to shove, if Congress really wants to, they can enshrine Net Neutrality into law. And it only took the Democratically led Congress in 2021, three weeks after the SCOTUS case to pass the new 300+ page law giving the EPA those new powers explicitly.
That’s the thing, the Republicans in the 118th Congress have shown they can not get anything done. They’ve pass 64 laws so far, most of them are renaming Post Offices and reupping funding to VA hospitals. They’ve spent almost 65% of the time in committee investigating various impeachment hearings. It’s so weird how they’ve had a majority in the House, could have worked on budget related things, and they’ve barely talked about the impending tax increase that’s coming once the tax cut act of 2017 runs out next year. They literally had planned to run on that sole thing back in 2017, that’s why they set it up to expire during an election year, and not a peep from them this year on it.
Meanwhile the Democrats in the 117th Congress passed 362 laws, with bangers like the CHIPs act, the Inflation Reduction Act, and the whole turn about is fair play with the whole EPA vs WV case. Because they took the majority they had and got things done.
So ISPs better hope Republicans can keep the mayhem up forever, because if Democrats do get into power in the House/Senate/and President. This whole stunt with the Supreme Court they’re pulling could massively backfire on them. Because if NN gets into law, well then it’s way harder to undo that.
I think the Israeli have not liked Palestinians for a bit longer than Biden has been anything.