Something Is Going Very Wrong With Google’s AI

Personally I BARELY if ever use genAI. But nano banana is noticeably worse now compared to in april and before. Thats not even mentioning it just makes shit up aswell. I have seriously no idea what they are doing over there, but it’s not looking good for them.

Google just have a different focus. Its stupid to compare them to coding-focused models.

Coding is just 1 thing among trillions (quite literally) of things they have to make cash. Our grandmas are using Gemini every day when they Google some basic question and read the AI answer it right there – it is incredibly useful and people underestimate it. Perplexity is trying that spending billions, Google just made a tiny adjustment to their interface.

Gemini is the only truly multimodal model – and it's fast, cheap, reliable. Send a video, an audio, whatever to it and ask anything. No other model can do it.

The AI race??? The Goldrush to oblivion, not a race you would want to be winning.

opino que google observa como anth y open se comen entre sí, para finalmente, entrar al final de la batalla y ser la vencedora
Like u e sai en la 2da GM 🔥

Google cooking. Mark my words.

a lot of time spend for bla bla bla?

3.7 flash in antigravity is quite good. The only problem with antigravity is the non-stop approve stream for everything. The model is fast, very fast and quite compotent. I like 3.7 flash a lot.

Yes 3.7 bare model seems like a flop but it is flash model. Why is it expected to be as good as deep thinking models. This is ehere your harness comes in. Dont forget gpp or google play Pokémon proved that exactly the flash model with a well structured self evolving harness beats fable on research and thinking and self evilving which was the base used with prime midel that beat fable on ark agi 3 benchmarks. So the thinking of 3.7 is there, just has to be steered correctly. After all you dont expect a small model to beat the big models outright, this is where the hitl collaboration comes in. Just food for though

Remember, Google just reported a loss for the first time since the early 2000s….why?

The reports of my death are greatly exaggerated

Maybe youre not seeing the same things I am seeing but I am please with them.

It's well known that AI will destroy their business model. They had to choose between destroying themselves or being destroyed by others, they chose the latter.

The latest updates have destroyed Gemini. It’s behind by about 6 months. They’re so bad it’s not even funny. Even when forcing some specific behavior it still goes on a ride of bullshit.

The mindless AI spin never ends. Do your research. AI can’t “do” maths, it can't verify its own results, it has no logic or deduction capability, it’s not trusted in law or medicine, it’s only right 50% of the time, and it can hallucinate. It added glue to a pizza recipe, it said the word Google had 2 "p"s, it said we should eat rocks for vitamins and minerals. It's the biggest scam in history. LOL. The AI companies and their clueless fanboys and influencers are fraudulent crooks and charlatans.

Probably just waiting to release new models when Googlebook releases. Also after Fables and models escaping, they probably refining and my last thing , I believe working on cost and efficiency is the future of AI.