Live data from Hacker News

Gemini AI

deepmind.google

111–120 of 1001 posts

Re: Gemini AI

#111
post #27

The improvement over ChatGPT are counted in (very) few percents. Does it mean they have entered a diminishing returns phase or is it that each percent is much harder to get compared to the previous ones ?

Training large language models is characterised by diminishing returns; the first billion training inputs reduce the loss more than the second billion, the second billion reduce the loss more than the third, etc. Similar for increases in size; the improvement is less than linear.

Re: Gemini AI

#112
post #44

What is up with that eval @32? Am I reading it correctly that they are generating 32 responses and taking majority? Who will use the API like that? That feels like such a fake way to improve metrics

Page 7 of their technical report [0] has a better apples to apples comparison. Why they choose to show apples to oranges on their landing page is odd to me.

[0] https://storage.googleapis.com/deepmind-media/gemini/gemini_...

Re: Gemini AI

#113
"We finally beat GPT-4! But you can't have it yet."

OK, I'll keep using GPT-4 then. Now OpenAI has a target performance and timeframe to beat for GPT-5. It's a race!

Re: Gemini AI

#114
Don't get me wrong, I'm excited to try it out.

I find it surprising that they only released Pro today, but didn't release the stats for Pro. Are those hidden somewhere else or are they not public?

Taking a different view on this release, the announcement reads, "We released a model that is still worse than GPT4 and, sometime later, we will release a model that is better than GPT4." which is not nearly as exciting.

Re: Gemini AI

#116
post #58

So, better than GPT4 according to the benchmarks? Looks very interesting. Technical paper: https://goo.gle/GeminiPaper Some details: - 32k context length - efficient attention mechanisms (for e.g. multi-query attention (Shazeer, 2019)) - audio input via Universal Speech Model (USM) (Zhang et al., 2023) features - no audio output? (Figure 2) - visual encoding of Gemini models is inspired by our own foundational work o…

That's for Ultra right? Which is an amazing accomplishment, but it sounds like I won't be able to access it for months. If I'm lucky.

Yep, at this point I'd rather they hold their announcements until everybody can access it, not just the beautiful people. I'm excited and want to try it right now, and would actually use it for a PoC I have in mind, but in a few months the excitement will be gone.

Re: Gemini AI

#117
post #70

it's really amazing how in IT we always recycle the same ten names... in the last three years, "gemini" refers (at least) to: - gemini protocol, the smolnet companion (gemini://geminiprotocol.net/ - https://geminiprotocol.net/ ) - gemini somethingcoin somethingcrypto (I will never link it) - gemini google's ML/AI (here we are)

Naming things is one of the two hardest problems in computer science, after all.

I don't even want to think about how much time I've wasted mulling over release names.

Re: Gemini AI

#118

They've reported surpassing GPT4 on several benchmarks. Does anyone know of these are hand picked examples or is this the new SOTA?

It will be SOTA maybe when Gemini Ultra is available. GPT-4 is still SOTA.

Usually SOTA status is established when the benchmark paper is released (probably after some review). But GPT4 is the current generally-available-SOTA

Re: Gemini AI

#119
post #17

Feels more like an Apple post "the best fastest blabla-est". How about making it available to try without the fluff?

I'm not a marketer but it's hard to see what the point of these glossy press releases without a call to action is. If I could have tried it today, I would have dropped everything and tried it. Now I will forget about it for a while and try it whenever I hear through osmosis that it's available. To the extent that I am excited and/or interested, the snooze button is pressed. Maybe that's the desired outcome?

I wonder what % of google shareholders understand this difference you talk about between “available now! try it” and “available Q2 next year”.

Re: Gemini AI

#120

Apple lost the PC battle, MS lost the mobile battle, Google is losing the AI battle. You can't win everywhere.

Beautifully said. So basically: Apple lost the PC battle and won mobile, Microsoft lost the mobile battle and (seemingly) is winning AI, Google is losing the AI battle, but will win .... the Metaverse? Immersive VR? Robotics?

Adblock war(?)
Post reply on HN