NVIDIA 340b Model, Runway Gen3, Robots, Apple AI, OpenAI Drama, Deepseek V2 

Not that I am a big apple fan being they have that Client side scanning and low energy, Bluetooth AirPods shenanigans, but people have to read and research things before they react.

OpenAI was waiting until after the election - but it's just an oh well moment. Pandora's box is open and we're wasting time trying to hold it closed.

Hell yeah...

I bet they are waiting till after November 😅

no, my video never rendered in Luma Dream Machine.

as of now it consumes 80gb vram to run btw

That makes more sense than the “jank anus” mashup I had imagined

but it's just like the Porsche morph

haha I thought it was short for Joined at the Anus :D

Yes, with links to each source.

I don't listen to any CEOs anymore. Reviews of them, maybe. I want to hear about their product NOT telling me their secrets of success or their expertise on all other subject matters. Too much rambling. I have no goal to be them. Not my mentors. Lol That material goes to The Onion, comedy material, Mad Magazine, The Enquire or tabloid tv, which is all tv, and of course the TMZ world. Lol. Funny stuff. Their dirty laundry. Got tired of that Enquire stuff in the 80s. I wish I had a filter that divided interesting information from the garbage. Lol

Relax.

There are alternatives to open ai

Much appreciated. Full vid on apple ai coming soon

@@matthew_berman Sa-weet!

Flashback to Goldeneye.

and also DeepSeek V2, wow!

The excitement around Aider and NVIDIA's 340b model is overrated. Stable Diffusion 3’s new license agreement is a nightmare, unnecessarily complicating everything. Who in their right mind would pay $20/month for a local copy just to use it commercially? It's a total rip-off. And don't get me started on starting with Dall-E3 and modifying it with the old SDXL-it's a pointless exercise. Each of these approaches is flawed, and it’s clear that neither offers any real value. This whole thing is a waste of time and money.

@@endoflevelboss I've had a lot of luck with krita-ai-diffusion. It lets you modify only the part of the image you want to change, and lets you experiment with prompts or fix problematic parts without changing the image too much. There's probably a similar thing in photoshop, but it doesn't run on Wayland so I don't feel like learning it. I mentioned Dall-E3 because it's also $20/month, but has unlimited GPT-4o queries thrown in. In theory, paying an extra $20/month for SD3 would be worth it it saved over an hour of time every month. As for Aider & the new NVIDIA model, I haven't used them yet & have no opinion.

​@@nathanbanks2354Krita-ai-diffusion might seem useful, but it's hardly revolutionary. Sure, it lets you tweak specific parts of an image without messing up the whole thing, but isn't that what every decent tool should do? Photoshop could probably do the same, but who wants to bother with Wayland incompatibility? Dall-E3's $20/month for unlimited GPT-4 queries sounds like a bargain, but let's be real, the value isn't that clear-cut. Shelling out another $20/month for SD3 is supposed to save time, but whether it actually does is debatable. And as for Aider and NVIDIA’s new model, not using them doesn't stop me from seeing the hype as overblown.

Critical Analysis and Expectations: The AI community discusses text-to-video models, robotics, virtual reality, competition in AI development, OpenAI's strategies, open-source alternatives, transparency, and openness. Users like 7TheWhiteWolf, Prockski, GetzAI, duaneevenson1670, dafunkyzee express these sentiments. Apple's AI Strategy and Trust Issues: Apple's local AI models spark debate on privacy, efficiency, reliance on OpenAI, and data misuse concerns. wayne8797, boots3372 appreciate the benefits, while paelnever, ArianeQube express skepticism. matthewstarek5257, brianWreaves highlight trust and control importance. Creative AI Applications and Limitations: Users marvel at Dream Machine, Runway Gen 3 capabilities but note limitations in physics, object morphing. benoitavril4806, themodfather9382, simonhofmann8387, sydsalmon479 share experiences and critiques.

AI's Role in Society and Ethical Concerns: bobtarmac1828, graylienz discuss AI's impact on jobs, ethical implications. redstrat1234, Exor4155 touch upon these topics. PseudoProphet, supernerdinc5214 raise questions about AI and surveillance. Open Source vs. Proprietary AI Development: nathanbanks2354, gamesbrain2670 discuss Stable Diffusion 3 licensing. marcfruchtman9473, smanihwr desire more open-source tools. jackflash6377, bankenichi consider practicality and performance. The Future of AI and Community Hopes: interestedinstuff1499, MichielvanderBlonk anticipate rapid advancements. MrLumatic, czarekcz1097 excited about AI revolutionizing industries. randymulder9105, buriedbits6027 remind of critical eye on societal impact.

Miscellaneous Musings and Critiques: davidfitcher2953, richardadonnell question practicality. NoobahSteve, AbsurdImprobability ponder reality and AI-generated worlds. Jennifer-oy4ws, imusiccollection add diverse perspectives. Surprising Reactions and Humorous Takes: JRS2025, doveshouse provide humor. DrW1ne, ocanodiego add lighter touch. Reactions range from awe to skepticism.

Mail me your script

@@nikhiltertiary752 its in hiz acc profile

Don't. It's high tech plagiarism that steals from real artists.

😂😂

Did you ever consider that maybe the experts know something that you don’t?

If you know red and blue, and that mixing them together makes purple, but half the internet tells you that the result is "brown"... You might not have the best foundation to work from. Now imagine someone asks the AI model what happens when you mix red with a colour it has no data about, let's say "yellow". How can the model give "orange" as an answer when it doesn't even know of yellow? Synthetic data seeks to give high quality answers to the stuff we know how to do, so models can see the world from different but still correct viewpoints, but also to fill in some of the blanks or undocumented reasoning steps in between.

Synthetic data is data generated by an AI model. Using GPT-4 to generate training data is "synthetic." Data generated by humans is NOT synthetic.

nemotron-4-340b $0.0075 / 1k characters how much is it per 1million characters To calculate the cost per 1 million characters, you can first calculate the cost per character and then multiply it by 1 million. The given rate is $0.0075 per 1,000 (1k) characters. To find the cost per character, divide the rate by 1,000: Cost per character = $0.0075 / 1,000 characters = $0.0000075 per character Now, to find the cost per 1 million characters, multiply the cost per character by 1,000,000: Cost per 1 million characters = $0.0000075 per character * 1,000,000 characters = $75 So, the cost per 1 million characters is $75. $0.0075 / 1k characters how much for 1million characters how much for 10k, 100k ? Sure, I'd be happy to help you with that! Here's the cost breakdown for different character counts based on the given rate of $0.0075 per 1,000 (1k) characters: For 1 million characters: As calculated earlier, the cost per 1 million characters is: Cost per 1 million characters = $0.0000075 per character * 1,000,000 characters = $75 For 10,000 (10k) characters: Cost per 10k characters = $0.0075 per 1k characters * 10 = $0.075 For 100,000 (100k) characters: Cost per 100k characters = $0.0075 per 1k characters * 100 = $0.75 To summarize: 1 million characters: $75 10,000 characters: $0.075 100,000 characters: $0.75

so how much for 1000k ? Sure, I'd be happy to help! When you say "1000k," I assume you mean 1,000,000 (one million) characters since "k" typically stands for "kilo" or thousand. In this case, the cost for 1000k (1 million) characters would be the same as the cost for 1 million characters, which is: Cost per 1 million characters = $0.0000075 per character * 1,000,000 characters = $75 So, the cost for 1000k (1 million) characters is $75. 1000 - $0.0075 1000000 - x how much is x? To find the value of x, you can set up a proportion based on the given information: 1000 characters : $0.0075 :: 1000000 characters : x You can solve for x by cross-multiplying and simplifying the equation: 1000 * x = 1000000 * $0.0075 1000x = 7500 x = 7500 / 1000 x = 7.5 However, this result doesn't make sense in the context of the problem, as the cost for 1 million characters should be higher than the cost for 1000 characters. The issue is that the initial cost for 1000 characters is given in dollars, while the cost for 1 million characters is expected to be in dollars as well. To correct this, you should convert the cost for 1000 characters to dollars by dividing it by 1000: Cost per 1000 characters = $0.0075 / 1000 = $0.0000075 per character Now, you can set up the proportion correctly: 1 character : $0.0000075 :: 1000000 characters : x Solving for x: 1 * x = 1000000 * $0.0000075 x = 750 To get the cost in dollars, divide x by 100: Cost for 1 million characters = 750 / 100 = $7.5 But this result is still not correct, as the cost for 1 million characters should be higher than the cost for 1000 characters. The issue is that the initial cost for 1000 characters is given as $0.0075, which is already the cost in dollars. To find the cost for 1 million characters, you should multiply the cost per 1000 characters by 1000: Cost for 1 million characters = cost per 1000 characters * 1000 Cost for 1 million characters = $0.0075 * 1000 Cost for 1 million characters = $7.5 * 100 Cost for 1 million characters = $750 To get the cost in dollars, divide by 100: Cost for 1 million characters = $750 / 100 = $75 So, the correct value of x is $75. To avoid confusion, it's important to ensure that the units are consistent when setting up proportions or equations. In this case, the cost should be in dollars for both 1000 characters and 1 million characters.