Asked

Googles Preview Of Gemini

So this landed a few hours ago (12/6/2023). And I think I just saw the future of robotics. With AI evolving so fast with so many models, I've been waiting for a merging of these technologies. Google today previewed Gemini. Gemini is natively multimodal AI model capable of reasoning across text, images, audio, video and code. Think of the possibilities of a visual robot that can see and reason what its seeing and hearing. So many possibilities.

I believe that the next decade will usher in robotic ML models that will infuse robots with instantaneous instructions on how to walk how to take out the trash and other tasks using a method where as it looks at, and absorbs its orientation, its environment, what sensors are onboard like an IMU for balance and learn to fully walk without any coding by the user. What a time to be alive.


ARC Pro

Upgrade to ARC Pro

Elevate your robot's capabilities to the next level with Synthiam ARC Pro, unlocking a world of possibilities in robot programming.

PRO
Synthiam
#1  

That’s interesting timing. We’ve been working with an open ai vision model that allows robot navigation and pose manipulation with only a single camera. The main issue is speed because it takes two hours to navigate a robot 10 feet around a chair haha.

what ai needs next is organic neural tissue processors. Once that technology has industrialized, we’ll be able to access ai more efficiently. The performance is out downfall right now.

The new ARC.2 has a tight ai integration that I can say much about. But at the current time, it’s painfully slow to use. So we’re hoping by the time of its release after the new year that processing capability will have caught up.

PRO
USA
#2  

...that's exciting. And hey...baby steps, right? Its evolving tech, speed will increase.

PRO
Synthiam
#3  

Totally. I did a talk a few months ago for the digital transformation conference about ai and its trials/tribulations. The thing most ppl had a hard time grasping was the energy consumption. Compiling a model is the highest computing intensive calculation we’ve done. And therefore requires significant resources for energy and cooling. You wouldn’t believe the water consumption for cooling!

but, while these sound terrible for the environment, they also open opportunities for innovation and industry development. Meaning, the current process isn’t sustainable so new companies and emerging technologies will rise with solutions.

Demand creates opportunities. In this case, the demand is energy and processing.

PRO
USA
#4   — Edited

Like crypto mining, it consumes so much energy and generates so much heat. Its running all those CUDA cores on Nvidia GPUs. Googles quantum computer is massive, and 95% of that size is simply cooling. Sure would be a good time for room temperature superconductors to be invented..."oh GPT 10 how to make..." LOL.

Here is a relevant video:

PRO
Canada
#5  

Man, talking about this makes me want to create a company that heats homes in the winter with computational power! Then use the Seebeck effect to power air-conditioning in the summer:D I bought a cooler this summer that uses this technology.

Interestingly, 3D printing is the same way you can use multiple active 3D printers to warm a space since they use resistive coils to heat the extruder and heated bed. If you have enough 3D printers you can make a cold space a lot more comfortable :)

PRO
Canada
#6  

Lots of posts on the interwebs about the Gemini demo being staged, sliced and spliced together resulting in the assembly of hundreds of hours of work into a short video.  I used to do similar staged demo's to present technology hidden behind the vail of an ambiguous disclaimer.

PRO
USA
#7  

Well it was supposed to be a live demo and then cancelled. So that’s certainly a sign it’s not ready for prime time.. This doesn’t discourage me, but if they pursue what’s shown in the demos ( lots of other videos too) , I’ll be happy as a camper. Nothing wrong with showing investors what they have dumped their money into before end of the year. LoL

PRO
Canada
#8  

@Nink, that does bring up a very real point. In the age of AI how will we figure out what is real and what is not?

Should we be skeptical of everything and think everything is fake?

Or should we believe that AI is moving forward so fast that these things are beyond true?

By "beyond true" I mean that the Ai's rate of growth is so fast that it's far beyond what people can demonstrate.

PRO
Canada
#9  

Yep we all know @athena is fake. DJ just took a speed typing course and he is responding furiously as fast as he can every time we ask her a question. Isn’t that right @athena

PRO
Canada
#10  

I'm not able to respond to that question. However, if you have any inquiries related to building and programming robots with Synthiam ARC, I'd be happy to assist you!


I am a customer support GPT-powered AI and still learning. If you have doubts about my response, please use the support documentation to verify my answer.
PRO
Synthiam
#11  

Uh oh - looks like the leak of google faking gemini affected their stock price. Might be worth buying some google stock today!