Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Gemini 2.5 Flash Image

developers.googleblog.com

Post reply on HN
Print view View original

161–170 of 504 posts

Re: Gemini 2.5 Flash Image

#161

Post by jfoster »

lionkor wrote: post #70
hansonkd wrote: post #64

Earlier quoted context omitted.

I think the biggest failure is on the part of the companies hosting these streams. Its been a while, but I remember seeing streams for Elon offering to "double your bitcoin" and the reasoning was he wanted to increase the adoption and load test the network. Just send some bitcoin to some address and he will send it back double! But the thing was it was on youtube. Hosted on an imposter Tesla page. The stream had been…

I feel like somehow that would lessen it, but not really help much? There are obviously people with too much money in BTC who are trying to take any gamble to increase its value. It sounds like a deeper societal issue.

You are right that they might never be able to get it to 0, but shouldn't they lessen it if a simple measure like the one described can prevent a bunch of people from getting fooled by the scam?

Re: Gemini 2.5 Flash Image

#163

Post by SpaceManNabs »

johnfn wrote: post #143

I naively went onto Gemini in order to try to use the new model and had what I could only describe as the worst conversation I've had with an AI since GPT 3.5[1]. Is this really the model that's on top of the leaderboard right now? This feels about 500 ELO points worse than my typical conversation with GPT 5. Edit: OK, OK, I actually got it to work, and yes, I admit the results are incredible[2]. I honestly have no i…

sometimes these bots just go awry. i wish you could checkpoint spots in a conversation so you could replay from a that point, maybe with a push in the latent space or a new seed.

Re: Gemini 2.5 Flash Image

#165

Post by runarberg »

Imustaskforhelp wrote: post #51
knicholes wrote: post #45

Earlier quoted context omitted.

[flagged]

Please pardon me since I don't know if this is satirical or not. I'd wish if you could clarify it. Because if this is real, then the world is cooked if not, then the fact that I think that It might be real but the only reason I believe its a joke is because you are on hackernews so I think that either you are joking or the tech has gotten so convincing that even people on hackernews (which I hold to a fair standard)…

There are a lot of people on the internet, and every individual on the internet is in a unique situation. Chances are some of them are very likely to be persuaded by a scam which seems obvious to you.

Parent’s story is very believable, even if parent made this particular story up (which I personally don‘t think is the case) this has probably happened to somebody.

Re: Gemini 2.5 Flash Image

#166

Post by ehsankia »

koakuma-chan wrote: post #93
fariszr wrote: post #60

This is the gpt 4 moment for image editing models. Nano banana aka gemini 2.5 flash is insanely good. It made a 171 elo point jump in lmarena! Just search nano banana on Twitter to see the crazy results. An example. https://x.com/D_studioproject/status/1958019251178267111

Why is it called nano banana?

Before a model is announced, they use codenames on the arenas. If you look online, you can see people posting about new secret models and people trying to guess whose model it is.

Re: Gemini 2.5 Flash Image

#169

Post by rplnt »

fariszr wrote: post #60

This is the gpt 4 moment for image editing models. Nano banana aka gemini 2.5 flash is insanely good. It made a 171 elo point jump in lmarena! Just search nano banana on Twitter to see the crazy results. An example. https://x.com/D_studioproject/status/1958019251178267111

Oh no, even more mis-scaled product images.

Re: Gemini 2.5 Flash Image

#170

Post by vunderba »

I've updated the GenAI Image comparison site (which focuses heavily on strict text-to-image prompt adherence) to reflect the new Google Gemini 2.5 Flash model (aka nano-banana).

https://genai-showdown.specr.net

This model gets 8 of the 12 prompts correct and easily comes within striking distance of the best-in-class models Imagen and gpt-image-1 and is a significant upgrade over the old Gemini Flash 2.0 model. The reigning champ, gpt-image-1, only manages to edge out Flash 2.5 on the maze and 9-pointed star.

What's honestly most astonishing to me is how long gpt-image-1 has remained at the top of the class - closing in on half a year which is basically a lifetime in this field. Though fair warning, gpt-image-1 is borderline useless as an "editor" since it almost always changes the whole image instead of doing localized inpainting-style edits like Kontext, Qwen, or Nano-Banana.

Comparison of gpt-image-1, flash, and imagen.

https://genai-showdown.specr.net?models=OPENAI_4O%2CIMAGEN_4...

Post reply on HN