Will have to test it but my last attempt with the gemini series API was disapointing Gpt 4o-mini was still way more interesting than Gemini flash as a small fast and cheap model And sonnet 3.5 than Gemini pro for a bigger cutting edge one
I'm shocked by how significant OAI's first mover advantage has been. Google literally invented the Transformer. AI studio is unquestionably the best price performance offering in the market. Gemini's context window unlocks use cases that were literally impossible with GPT. I don't agree that people who haven't coded can tackle any code problems using coding models. LLMs are human level prompt engineers; however, programmers are much better at prompting coding models than non-technical folks. Agreed that people are focused on the future more than what current models are capable of. We've just seen the tip of the iceberg. Kurzweil thinks BCIs for the final form factor.
Benchmarks suggest the model can manage to complete something. Last I checked ‘Pro’ 1.5+ couldn’t lead an experienced Google console user around its basic product offerings, instead it itself recommending Azure. Note this is the paid ‘enterprise’ version that exists to assist with cloud console.
All outlined in the official blog post here: developers.googleblog.com/en/updated-production-ready-gemini-models-reduced-15-pro-pricing-increased-rate-limits-and-more/
We were expecting Multimodal Input Multimodal Output hugely capable Gemini 1.5 Ultra that can also do inference time compute! Not some mere improvements!!
So disappointed with the big tech companies, no matter how much money they pour into PR i will never accept their monopolization of AI and by extension humanity's future.
Open AIs advanced voice is actually pretty bad though🗑️ 😂 It gives up when I ask it to sing for me and starts yelling at me randomly. Like literally yelling lol 😆 It’s really buggy and crazy 😜 especially when you use it with other languages too. It isn’t able to slow down like in the demo or do near as good as it did in those demos 😂but maybe it’s still getting tweaked hopefully 🤞
Nice talk. Good to see Gemini getting better.
I'm in the 1% Yahoo...!!
Atleast, that's the one positive feedback I got
Cheers Rowan
Will have to test it but my last attempt with the gemini series API was disapointing
Gpt 4o-mini was still way more interesting than Gemini flash as a small fast and cheap model
And sonnet 3.5 than Gemini pro for a bigger cutting edge one
How good is 1.5pro002 compared to pro experimental 08/27 on ai studio
This was not the release I was expecting. Very excited for it.
Gemini has really improved its docs for ease of use. Looks like Logan was behind that improvement for the dev facing stuff.
Gotta love that 2M context. Nobody is even close.
Early adoption indeed. Thank you for these efforts
The responses reflect largely the usability of AI based on the deep (or, lesser clear knowledge) of the problem’s problem to be solved via AI!
I'm shocked by how significant OAI's first mover advantage has been. Google literally invented the Transformer. AI studio is unquestionably the best price performance offering in the market. Gemini's context window unlocks use cases that were literally impossible with GPT. I don't agree that people who haven't coded can tackle any code problems using coding models. LLMs are human level prompt engineers; however, programmers are much better at prompting coding models than non-technical folks. Agreed that people are focused on the future more than what current models are capable of. We've just seen the tip of the iceberg. Kurzweil thinks BCIs for the final form factor.
Still waiting for Gemini Live
Rowan do you have any Japanese Background ?
Where are the benchmarks ?
Benchmarks suggest the model can manage to complete something. Last I checked ‘Pro’ 1.5+ couldn’t lead an experienced Google console user around its basic product offerings, instead it itself recommending Azure. Note this is the paid ‘enterprise’ version that exists to assist with cloud console.
All outlined in the official blog post here: developers.googleblog.com/en/updated-production-ready-gemini-models-reduced-15-pro-pricing-increased-rate-limits-and-more/
Humanoids is the final form of agentic...
I wonder who is hitting him up about Google models other than to make fun of
Funny things they both have profile pic on there account yet I haven't seen there face . Is this normal?😅
Google continuining to trail behind.
🇧🇷🇧🇷🇧🇷🇧🇷👏🏻
We were expecting Multimodal Input Multimodal Output hugely capable Gemini 1.5 Ultra that can also do inference time compute!
Not some mere improvements!!
Google took the L today
Google sucks.
Yes when I used Gemini it was so inaccurate.
But Dude ones you customize it it's Killer.
Google is actually really good.
Big time! They need to dry up the swamp right now or they are out.
+ 🎉
Just their documentation.... Also the separation between vertex and the other Google Gemini SDK these are really annoying......
So disappointed with the big tech companies, no matter how much money they pour into PR i will never accept their monopolization of AI and by extension humanity's future.
I don't think you have a non technical audience
What about all the woke propaganda? 🤔
Google AI is trash
Gemini update 🪫
Open AI’s Advanced Voice drop 🔋
Open AIs advanced voice is actually pretty bad though🗑️ 😂
It gives up when I ask it to sing for me and starts yelling at me randomly. Like literally yelling lol 😆 It’s really buggy and crazy 😜 especially when you use it with other languages too. It isn’t able to slow down like in the demo or do near as good as it did in those demos 😂but maybe it’s still getting tweaked hopefully 🤞