Well ya, exactly. But many other companies dumped billions of dollars into it, most gave up, and a few are still struggling. I think my points stands, no matter how much of a snake oil salesman Musk is other companies believed this was solvable with a bit of money and time.
Whelp, here we go. Constant screenshots of your computer displays can't possibly go wrong . . .
I actually used this kind of program for quite a while when I was doing contract work.
Along with periodic screen shots it also recorded the name of the application I was in and the name of the file I was editing!
I used it to keep track of hours worked for billing, it worked really really well. Made it super easy to see "I started working in this project at 8:45, took a break at 10, started again at 10:20, etc..." I had the screen shots in case my client wanted some kind of evidence (none ever did), but a few times it was helpful in going back to recreate something I'd done but neglected to document.
So if if it's off by default, only stores information locally, and can be heavily and separately encrypted then it could be useful for some people.
But I haven't even read anything about its intended use and I'll take a guess that it's not like that (though I did read they're going to set it off by default which is the bare minimum).
EDIT: And I could turn it on and off as I wanted, though after a while I just left it on and figured if someone could get and decrypt the file they deserved to see which ecchi anime I watched.
__________________ Uncertainty is an uncomfortable position.
But certainty is an absurd one.
A buddy of mine was wondering the other day if for financial and emotional health reasons polygamy and polyamory will begin to skyrocket. Like a controlled open relationship with additional relationship, household management and other benefits.
Apparently it's a concept called a throuple.
I have noticed an uptick in articles on polyamory in places like the New York Times. I think it may be a future front in the culture wars. Wonder what the flag will look like…
The Following User Says Thank You to edslunch For This Useful Post:
Yes, noted AI expert *checks notes*, Pope Francis is going to tell the world what his perverse moralizing has to say on the subject. This should be a big help in guiding humanity.
Why is everyone pushing their AI searches? I don’t think I need it for searching text in Adobe.
Because it's the next level of buzzword expectation?
Device -> "Smart" device -> "Smart" device "with AI technology" (We are here) -> "Smart AI" device -> Device (Maybe... if the world hasn't been destroyed by then)
Remember how everyone tried to crypto-currency and NFT everything? Ya, the next fad cycle.
Spoiler!
We are somewhere on the first rise, probably not even halfway. Expect a lot more.
I don't know... we've got groups around the planet scrambling for power resources and people floating the idea of $100 billion dollar computer clusters. Sam Altman is actively exploring the possibility of starting his own nuclear power plants. This is only half-way up the hype curve?
You could be right though. I think crypto is a solid example of a hype folly. That said, perhaps it's just not time for crypto yet?
Quote:
Originally Posted by Wormius
Why is everyone pushing their AI searches? I don’t think I need it for searching text in Adobe.
To me a successful search means the AI can manage a large amount of something. Ya, it looks like expensive Command-F, but once it can hold something the length of a book in memory, it can start doing more and more complex tasks. How many jobs could be distilled into something smaller than a Harry Potter book? That's where things start to get pretty bumpy.
I more meant halfway up the hype curve, in that it hasn't been shoe horned into everything yet, and the first AI crash is a ways off. If we were near the peak I'd expect the crash in under 6 months, but I think we are farther out than that. My dishwasher doesn't even have AI yet!
Your best bet is Stable diffusion (1.5 or SDXL models) if you able to run it locally on your computer but it requires a decently beefy computer.
Microsoft co-pilot uses Dalle-3 for image generation (the most coherent generator but least realistic) where you get a few a day. OpenAI doesn't allow dalle image generation for free (just check) but it comes with chatgpt subscription and is a great option.
There are the odd options such as mage.space which offers limited free SFW image generation using SD, and a few apps that do as well. Meta has their own offerings with free image generation if you want to play with something.
The Following User Says Thank You to Firebot For This Useful Post:
You'll want to keep your expectations in check also, as MidJourney is currently king of realism. Dalle-3 is cool, but it still has that cartoony look most of the time that is most often associated with an AI feel. If that's all you're after, have at it.
The Following User Says Thank You to Russic For This Useful Post:
There has been some absolutely crazy advances in AI models in the past month that may have gone unnoticed.
Sonnet 3.5 was silently released a few weeks ago by Anthropic which has beaten gpt4-o on all tests except math, but the biggest new advantage is how powerful it is at coding. And this is their mid tier model. I've been using it via Perplexity and although I don't have access to Artifacts this way to optimize use, but the coding is significantly improved with way less errors than anything prior It made a fully working enhanced Flappy Bird game with the same physics within 3 prompts as a quick example.
Luma AI released Dream machine last week which went quite viral considering how out of nowhere it came, giving near Sora level of video to the public for free. The best use is image to video which keeps high fidelity to the original image, Feed it a Midjourney image or a normal photo for best results. Free generations are now limited to 5 a day and 30 a month, with 150 available for the first tiered level.
Yesterday Runway AI released Gen 3 Alpha to the public, which offers even better near Sora like video. At the moment unfortunately image to video is not yet available to paid users (the non-public alpha version had it which was mind blowing) and it's quite expensive considering the limited generation. I am currently paying for the Pro tier, which gives approximately under 4 minutes of video generation for 35$ USD (Luma Dream Machine is currently much better value with image to video enabled). Since I am planning commercial use I am likely going to upgrade but hoping for image to video to be released before proceeding.
Maybe I can take a prompt or two and post a video of it via streamable to show its capabilities, and link a completed 1 min video I am working on. Video uses a combination of Runway, Midjourney (will be used for reference as soon as image to video is available), Suno for music and Elevenlabs.io for voice and sound effects, video prompts generated through Sonnet 3.5 with Runway custom prompting guide.
