Hey ChatGPT, Summarize Google I/O

Author Avatar

WVFRM Podcast

Joined: Mar 2024
Spread the love

Hey ChatGPT, Summarize Google I/O


This was a week full of AI events! First, Marques gives a few thoughts on the new iPads since he missed last week and then Andrew and David bring him up to speed with all the weirdness that happened during Google I/O and the OpenAI event. Then we finish it all up with trivia. Enjoy!

Chapters
00:00 Intro
01:17 Marques iPad Thoughts
16:49 OpenAI…

source

Reviews

0 %

User Score

0 ratings
Rate This

Sharing

Leave your comment

Your email address will not be published. Required fields are marked *

39 Comments

  1. I'm a big fan of the podcast and always enjoy the discussions. In this episode, the conversation seemed to focus more on the drawbacks than the potential benefits of the topic. I appreciated when MKBHD chimed in with a different perspective. It would be great to hear more about how others might interpret or use this information in future episodes. Perhaps you could invite a guest with a different viewpoint to offer a balanced discussion.

  2. Not just a review approach, you also seem to need a different take on podcast talks. It’s just getting boring to listen to the same conspiracy theories.

  3. It's weird how these tech guys don't understand just how much people use AI for. It's like they don't understand that a lot of people use these tools to write a lot of stuff that isn't too important but they need to, eg "write an email to the professor telling him I missed my class and want a retake" etc.

    There are also a lot of people using it for material and ideas. Ofcourse people are gonna ask about jokes, ideas, and other stuff to prepare for potential meetings. It's just a way quicker way of getting some good ideas instead of just sitting and thinking. You can get the prompts and then think with the added information.

  4. Andrew is, like, blah blah, like, blah blah, like, so like, like like like like like like, so like, go like. Speak goddam English, buddy. Maybe AI can say a sentence without throwing in 15 likes.

  5. Google Gemini is probably better for me, working in a Google school (though I'd need to use my personal account to access the tools). However, nearly all the features like Veo or Imogen 3 are not available in the UK (or the rest of the EU), so here I come ChatGPT.

  6. I think copilot is WORSE than chat gpt… idk it feels like it added a "IM SORRY DAVE" layer to chat gpt, where chat gpt doesnt give a f about what you ask, it will answer, and the answers are pretty good

  7. Andrew there are people at Lowe’s that can answer questions. Ellis the most American measurement unit is Cheesecake Factory menu (it shows your economic background).

  8. I feel like Marques is mentally checked out. Also I’d love for you to talk to someone about why llms aren’t being bonded with assistants.

  9. I don't think OpenAI did a good enough job of conveying how important and impressive it is to have a natively multimodal model that works this well. 4o is a very significant accomplishment. It's far ahead of GPT-4 in capabilities for half the price, and its design is far more scalable. Their event marked a major step forward for the field.

  10. @22:20 Come the hell on MKBHD let's give it props for what it can do, You sometimes forget it's still EART with Stupid Humans who would find something that intelligent a threat let's ease into it. For where it's at if it's facts it's pretty amazing i only say that because i don't know if she will sound like that on my phone or computer. Take it easy mr.nissan has never made a sports car lol

  11. Google has to be really careful, and they are, but they're definitely ahead of OpenAI. The massive context window and low API costs alone are enough to demonstrate that.

  12. Would love for y’all to mention, even for a couple minutes, the biggest drone company in the world, Dji, facing a potential ban. I know y’all don’t talk politics but more people need to hear about this.

  13. The content of the information itself is as important as the source. Even if this works well and answers your question perfectly, if it doesn’t allow you to change its sources or at least give you a couple of different responses based on different ones, it’s going to increment disinformation wildly. And don’t get me started on top results with the current state of SEO…

  14. The book club idea is on point. I've had the same desire for years now. I want to talk about the book but with someone who's experiencing the book at the same time and possibly same pace as I am.

  15. I think the lack of "energy" on Google side was because of Chgpt40 demo the day before. It already did almost all the stuff Google was presenting with "coming soon" and with a wittier, funnier, "more humane" interface… The marvel tech progress they had to show up, was just trounced the day before.

  16. I see Gemini evolving to be of more practical use by adapting it to YouTube and help channel creators create their brand, intro videos, logo, credits videos, and even generate music and visual content to add to their channel. All without any copyright infringement, if not AI education guiding channel creators towards the legal and right path to create post-able and shareable content and monetize on their channel.

  17. We have custom Copilots at the company I work at. It does make sense because we feed each Copilot with the files of the specific department.

  18. I felt the need to commen this because of all the AI haters, I own a google pixel phone and I use the recorder app to record all my college classes so that I don't have to take notes, I then export it to google drive and import into NoteBookLM, and use NoteBookLM as a tutor for all of my classes. Because all of the information comes from the recording transcription from my pixel NoteBook LM doesn't hallucinate a at all and I literally used it on a daily basis allowing me to take fast-track honors courses. I'm now eligible for scholarships which I wouldn't have had if I didn't use NoteBookLM. NoteBookLM's only limitation is math. The ChatGPT Store/Explore page has custom built mathGPT's that are designed to do math and don't hallucinate when asking math questions. I used their tutoring and generations of practice problems to maintain a 90% in my math class which would have been literally impossible to do because I live with my retired grandparents who are immigrants who can't do college level math and my parents are busy working so they can't tutor me. The only limitation for the MathGPT's is graphs and charts which they can't process so I used khan academy to learn about those. ChatGPT-4o's multimodality is extremely important because if it is done properly it should theoretically be able to interpret graphs, charts, and tables and explain them saving me (personally) hours of trying to learn math from YouTube and is a big step for an equal opportunity education for people who don't have anyone who can teach them at home.