Google Astra: O Futuro dos Assistentes de IA – Novidades e Inovações

This is prerecorded not a live demo

Lol… Cute that they had to say "This was captured in real-time". At least they're learning their lesson about that finally. By now everyone knows the original Gemini demo was mostly smoke and mirrors. But I'm still kind of skeptical, they really don't have the best track record with their AI marketing. Do this again in front of an audience. I can push TTS to my phone with the camera app open too, that's not proof of anything. Google does have great AI tech, but this seems desperate.

I want Google to do well, OpenAI needs competition. But this doesn't seem like an actual product so much as a product concept. Have they at least given journalists some hands-on time with it?

And Google will likely prevail in the Video Generation market, they have YouTube for training data, more data than any other competitor could muster. They'll likely beat SORA in the long run. But this is underwhelming. OpenAI demoed similar functionality yesterday, but their presentation was more convincing and seems to be significantly further ahead, and we already have access to test those features as they were demoed (save for the app on Windows / Android). But that app wasn't what was impressive, it's just an interface.

For example, pointing a camera out a window and asking "What town am I in"… They pulled that from the GPS data on the phone, not by analyzing the images. They would be stupid if they didn't, phones already have this functionality and have since 2007, and referencing the GPS data is orders of magnitude less processing intensive. If they did use AI for that, it's tremendously inefficient. Of all the questions to ask AI, why choose one that doesn't require AI to answer?

"Where are my keys"… Yeah, I could do that years ago with an Intel Movidius dongle, a camera, and a basic TTS voice. That's not new either. It's a cool concept that could help people in day to day life, I appreciate the thought, but that's not impressive from a tech standpoint though. Just a solution looking for a problem.

The code analysis… They're pretty good at that. I'll give them that one, but it would have been better if we could see the code in the video. For all we know it was a recipe for chocolate chip cookies on the screen with some odd indenting, lol. Not saying it was, but that's going to draw skepticism, at least until lots of users have time with the model to test it. It's not a dramatic step up from what exists already, these models can already describe code, they just added some OCR.

The whole video seems hastily thrown together, like "Oh Shit, OpenAI just beat us to the punch, here's the features they showed off, do that" and it was hurriedly cobbled together at the last minute. I might be wrong, I hope I am. But nothing demoed here is beyond the capacity of a hobbyist to build for themselves. I like the concept, the vision they have for this, but it doesn't feel like we're looking at actual tools. This is to keep shareholders off their ass, lol.

Veo: Google's NEW Text-To-Video AI Model! Sora Alternative!: https://youtu.be/qTRtLSiKtBY

I'll be honest, this is not as impressive as GPT-4o, especially when considering that OpenAI is including that model in the free tier

💗 Thank you so much for watching guys! I would highly appreciate it if you subscribe (turn on notifcation bell), like, and comment what else you want to see!

📆 Book a 1-On-1 Consulting Call WIth Me: https://calendly.com/worldzofai/ai-consulting-call-1

🔥 Become a Patron (Private Discord): https://patreon.com/WorldofAi

📖 Want to Hire Me For AI Projects? Fill Out This Form: https://td730kenue7.typeform.com/to/WndMD5l7

🚨 Subscribe to my NEW Channel! https://www.youtube.com/@worldzofcrypto

🧠 Follow me on Twitter: https://twitter.com/intheworldofai

Love y'all and have an amazing day fellas. ☕To help and Support me, Buy a Coffee or Donate to Support the Channel: https://ko-fi.com/worldofai – Thank you so much guys! Love yall!