Marvin
About
- Username
- Marvin
- Joined
- Visits
- 127
- Last Active
- Roles
- moderator
- Points
- 6,920
- Badges
- 2
- Posts
- 15,551
Reactions
-
Who will deliver useful AI to the masses?
mpantone said:AppleInsider said:I think it's too late to take any bets on who is going to drive mass market adoption of AI/ML tools.
Copilot for developers (Github Copilot) has integrations directly into editors, can rewrite parts of code and works very well.
Midjourney and Leonardo AI have millions of paid users and have generated hundreds of millions of high quality art images.
Adobe has some art tools baked into different apps (1:05):
Audio AI like ElevenLabs and Respeecher can simulate voices very accurately and movies are using these tools to use iconic voices in new productions:
https://techcrunch.com/2022/09/26/ai-is-taking-over-the-iconic-voice-of-darth-vader-with-the-blessing-of-james-earl-jones/
https://www.theverge.com/2022/12/1/23488336/disneys-ai-tool-aging-actors-vfx-announcement
There won't be a single AI provider that excels at everything. AI covers a wide range of fields and there will be top providers in each field.
AI tools are production-ready today, being used by millions and generating billions in revenue. -
Surgeons say Apple Vision Pro saves them pain and injury
hogman said:All I got from this is, how does a monitor cost $30,000? Especially when all the monitors I see in hospitals display simple graphs and numbers.
https://www.monitors.com/collections/surgical-displays
https://www.monitors.com/collections/surgical-displays/products/barco-mdsc-8358-k9307938
https://synergymedco.com/product/sony-lmd-xh550mt-55in-4k-3d-2d-lcd-medical-monitor/ (medical features listed)
https://www.medicalecart.com/products/sony-lmd-xh550mt-55-inch-4k-3d-2d-lcd-monitor-high-performance-medical-monitor-box-of-01.html
Low glare, picture-in-picture, 3D image input, color accurate, designed for easy cleaning for hygiene etc.
-
Commemorating Steve Jobs and his continuing influence on technology
nubus said:charlesn said:As for Vision Pro being a dud--that's hilarious for a product that has been in the hands of consumers for all of 32 weeks. Please cite fact-based sources for its "failure" that include Apple's actual internal projection for sales and how actual sales have fallen short of that number.There are Apple patents for this product from 2006:The images in the patent look almost exactly like AVP.The problem is this technology is still at the large helmet stage, even after nearly 20 years since this patent was filed and the above video was made.
What it really needs is a major technology breakthrough like how the multi-touch glass interface enabled the iPhone. There's a missing piece for AR and they are stuck with trying to shoehorn technology that's currently available to make it work. -
macOS Sequoia can run on Valve's Steam Deck with hacks
tipoo said:Now if only Apple dropped a bag of cash at Valve to bring that ARM Proton port to macOS natively, so that Steam games "Just Work" on macOS...That's the dream.
Valve's Proton has a DirectX to Vulkan translation layer (DXVK), which would need MoltenVK to convert to Metal as Mac doesn't support Vulkan directly.
Crossover and Whisky integrate Apple's D3DMetal translation to run Windows games like Proton. It supports a lot of games, there are over 500 games tested on the following channel:
https://www.youtube.com/@macprotips/videos
It supports higher-end recent games like Horizon Forbidden West:
The only problem with this setup is you have to install Steam inside the compatibility layers, then install the games inside this, sometimes with patches. It's not very user-friendly.
If it was possible to use either the App Store or native Mac Steam and install a pre-wrapped version of each game that is already setup and tuned, that would be much easier and a license per install could be paid to the company that provides the wrapper if it's from a 3rd party. -
Apple is reportedly not investing in OpenAI
chasm said:The "Harry Potter" models are interesting, but I don't quite understand why "Harry" always looks freshly beaten up in the graphic provided.The AI model training works like search engines where they scan images and then add labels. This is why high quality training data is needed for good models. The marketing material for Harry Potter shows images like this were he is injured:The AI does face recognition on images like this to find it contains an image pattern for Harry Potter and stores the pattern. When someone asks for Harry Potter, it uses this pattern as a source. It may not understand that this pattern contains injuries but AI models will eventually be trained to understand this. For now, people have to use negative prompts to tune image output so they request that the output doesn't match certain patterns.