Learning Library

← Back to Library

Google's AI Surge Dominates 2024

Key Points

  • Google dramatically shifted the AI landscape by unveiling nine new products in a matter of weeks, outpacing OpenAI, Anthropic, and AWS and silencing the narrative that it was still “catching up.”
  • The company launched Gemini 2.0, a state‑of‑the‑art language model so fast that developers are asking it to throttle its output because the streaming text is breaking downstream applications.
  • Google introduced Willow, a quantum‑chip prototype that demonstrably reduces errors as more qubits are added, marking a major step forward in practical quantum computing.
  • Its new Video FX (also called V2) outperforms OpenAI’s Sora, handling complex multi‑part prompts such as a chef cutting steak with realistic steam, motion, and textures.
  • Image FX was released as a “MidJourney killer,” sparking buzz that the tool could dramatically disrupt the current leader in AI‑generated imagery.

Full Transcript

# Google's AI Surge Dominates 2024 **Source:** [https://www.youtube.com/watch?v=0ZXTTZQolso](https://www.youtube.com/watch?v=0ZXTTZQolso) **Duration:** 00:06:46 ## Summary - Google dramatically shifted the AI landscape by unveiling nine new products in a matter of weeks, outpacing OpenAI, Anthropic, and AWS and silencing the narrative that it was still “catching up.” - The company launched Gemini 2.0, a state‑of‑the‑art language model so fast that developers are asking it to throttle its output because the streaming text is breaking downstream applications. - Google introduced Willow, a quantum‑chip prototype that demonstrably reduces errors as more qubits are added, marking a major step forward in practical quantum computing. - Its new Video FX (also called V2) outperforms OpenAI’s Sora, handling complex multi‑part prompts such as a chef cutting steak with realistic steam, motion, and textures. - Image FX was released as a “MidJourney killer,” sparking buzz that the tool could dramatically disrupt the current leader in AI‑generated imagery. ## Sections - [00:00:00](https://www.youtube.com/watch?v=0ZXTTZQolso&t=0s) **Untitled Section** - ## Full Transcript
0:00the story of the year is Google and 0:02their AI catchup everything else is 0:04second 0:05place Google started the year as it has 0:08been in 2024 I did not spend much time 0:12talking about what Google was doing 0:14because Google was not doing very much 0:17open AI was the story openi was 0:20launching stuff anthropic was launching 0:22stuff AWS was launching stuff at 0:25reinvent and now Google has come and 0:28trumped all of that in a matter of a 0:30couple weeks and I just blown away this 0:33was supposed to be open ai's 12 days of 0:35open AI they wanted to take Center Stage 0:38here they so far have really failed even 0:41though they dropped 01 which is an 0:43incredibly powerful reasoning agent well 0:47not quite an agent right it's an 0:49incredibly powerful reasoning llm and it 0:51will probably be an agent as soon as 0:53they integrate it into an agent 0:54framework in January but for now they're 0:58not the center of attention I want to go 1:00through nine releases that Google has 1:03launched in just the last few days 1:06because I want you to remember how much 1:09Google is shipping to change the 1:11narrative that Google needs to catch up 1:14number one Google shipped a new language 1:16model Gemini 2.0 it's absolutely 1:18incredible it is a state-of-the-art 1:19language model it's incredibly fast it 1:22is so fast they are having to ask the 1:26flash model to slow down people are 1:29writing into Google saying this whole 1:31idea of streaming text is breaking 1:32because Google Gemini 2.0 flash is too 1:36fast it's too 1:38fast imagine asking your product to slow 1:40down I've never had that happen they 1:43launched Willow a Quantum ship that 1:45shows error reduction as you add cubits 1:47which is an absolutely massive 1:49development they launched video FX that 1:52was just yesterday video FX is hands 1:57down better than Sora right now 2:00and people are being shy about that but 2:02I won't be because I have seen the video 2:04comparison if you pick the correct 2:07prompt a prompt that requires multiple 2:09parts of the video to work in parallel 2:11like a real world model does it is clear 2:15which one actually understands the real 2:17world and can produce good video better 2:19and it's video FX they also call the 2:22model itself V2 ve2 so if you see that 2:24floating around it's the same model 2:26here's an example prompt that I saw 2:30a chef cutting a piece of steak with 2:32steam rising from the piece of steak 2:35sounds simple easy to describe but if 2:37you have to visualize that with a video 2:40you have to have the knife moving 2:41correctly through the meat the texture 2:43needs to feel right so the knife is 2:44cutting appropriately not too easily you 2:47have to have the piece of meat moving 2:49just jiggling slightly in the right way 2:51in parallel with the knife as you saw at 2:53it you have to have the steam rising 2:55from the right parts of the meat clearly 2:58you have to have the hand working 2:59correctly with the knife there's a lot 3:02the only one I saw that worked was video 3:08FX and look that was just release number 3:10three there's nine of these image FX 3:14it's a mid Journey killer it's a mid 3:16Journey killer and you know it is 3:17because mid Journey's founder went on 3:18Twitter and got very depressed and 3:21defensive about how he thinks he's 3:22running a great company which by the way 3:24I love Med journey I still use M journey 3:26I do think it's a great product so 3:28nothing against M Journey but for casual 3:31users who want to generate 3:33images the tools that Google is 3:36launching are just not even close like 3:39they're so much easier you don't have to 3:40go to a Discord with image FX you just 3:42go to Google labs and you start playing 3:45around and you type and you see images 3:47it's what it should 3:48be and I'm not even getting to whisk 3:51whisk wasn't on my list of nine whisk is 3:52a cute little play toy for kids where 3:55you upload a couple of photos and you 3:56put them together and you can make like 3:57a sticker looking thing right like they 3:59just threw that out there like hey we've 4:01also got 4:02this music 4:04FX there's also a DJ product that they 4:07dropped I haven't even had time to play 4:08with it it's like what I do to play with 4:10AI stuff and Google's launching so much 4:12I don't even have time for it deep 4:14research absolutely incredible I do wish 4:16they would increase the token limit on 4:18this it will survey so many different 4:21sites in a short period of time it it 4:23literally Cuts research Time by tens of 4:26hours it's like a 10x saver on Research 4:28it's accurate citation is going to 4:30transform Academia in 4:322025 and the only thing stopping it from 4:35being a complete Banger is that it 4:36limits you to about five pages of output 4:38I think they're going to fix that it's 4:40it's 4:41amazing that was the feature that made 4:43me sign right back up for a paid Gemini 4:46subscription I needed the Deep research 4:49notebook LM plus so they launched 4:51notebook LM now they're refining it new 4:53interface chance to talk with the host 4:55notebook LM 4:56plus they're they're not just shipping 4:59and forgetting anymore they're actually 5:00shipping 5:02improvements 5:04Astra so Astra is a situationally aware 5:07AI agent where you can go in and say Hey 5:10you talk to your phone you talk to your 5:11Google pixel and say hey uh what are 5:14what are the lanterns hanging on the 5:15temple up there and as will answer or 5:18hey what does this sign say in Japanese 5:21uh I'm going to the train station and it 5:23will tell you if it's going to the train 5:25station or not that's what Astra is for 5:26it's that sort of situationally aware 5:28look through the camera tell you what's 5:29going on 5:30by the way Gemini is using Astra I think 5:34with the uh live streaming feature where 5:36it live streams off your laptop or off 5:37your phone and you can actually see 5:39what's going on and talk to Gemini very 5:41similar thing so it launched Astra 5:43that's for consumers and then it also 5:44launched Mariner which is agentic 5:46browsing like book my tickets when I 5:48take my next 5:50trip all of that in the space of just a 5:53few days Google is back and I think that 5:57one of the things that's going to be a 5:58major storyline in 202 6:01is with Google's momentum where it is 6:04are the capital reserves that Google 6:06brings to the table enough to enable it 6:08to win the market in 2025 in a way that 6:10open AI has been unable to do they 6:13haven't raised the funds for their GPT 6 6:15training run they are rumored to be 6:16having trouble doing that if they run 6:19into funding 6:20issues that could have ramifications for 6:23Microsoft's whole position in the AI 6:25space it's really really interesting it 6:26doesn't mean Microsoft is done by any 6:28means don't count them out 6:30but Google is shipping in a way that 6:32none of the other major players 6:34are so congrats whoever whoever's 6:37shipping at Google well done very 6:40excited love the quality um and I'm 6:42excited to see what's ahead cheers