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

Gemini 2.0: our new AI model for the agentic era

blog.google

Post reply on HN
Print view View original

481–490 of 512 posts

Re: Gemini 2.0: our new AI model for the agentic era

#481

Post by boringg »

StableAlkyne wrote: post #152
crowcroft wrote: post #15

Big companies can be slow to pivot, and Google has been famously bad at getting people aligned and driving in one direction. But, once they do get moving in the right direction the can achieve things that smaller companies can't. Google has an insane amount of talent in this space, and seems to be getting the right results from that now. Remains to be seen how well they will be able to productize and market, but hard…

> Remains to be seen how well they will be able to productize and market The challenge is trust. Google is one of the leaders in AI and are home to incredibly talented developers. But they also have an incredibly bad track record of supporting their products. It's hard to justify committing developers and money to a product when there's a good chance you'll just have to pivot again once they get bored. Say what you w…

Can I add they have a bad track record of supporting new products. Gmail, google, gsuite seem to be well supported.

Re: Gemini 2.0: our new AI model for the agentic era

#482

Post by echelon »

nilayj wrote: post #11
echelon wrote: post #7

Gemini in search is answering so many of my search questions wrong. If I ask natural language yes/no questions, Gemini sometimes tells me outright lies with confidence. It also presents information as authoritative - locations, science facts, corporate ownership, geography - even when it's pure hallucination. Right at the top of Google search. edit: I can't find the most obnoxious offending queries, but here was one…

can you provide some example queries that Gemini in search gets wrong?

I just found another one:

> A depsipeptide is a cyclic peptide where one or more amide groups are replaced by ester groups.

Depsipeptides are not necessarily cyclic, and I'd probably use "bond" instead of "group".

https://imgur.com/a/YslvJO2

These errors are happening all the time.

Re: Gemini 2.0: our new AI model for the agentic era

#483

Post by nuancebydefault »

zamadatix wrote: post #456
internet_points wrote: post #414

Earlier quoted context omitted.

I got so hopefuly by your comment I showed it my current bug that I'm working on, I even prepared everything first with my github issue, the relevant code, the terminal with the failing tests, and I pasted it the full contents of the file and explained carefully what I wanted to achieve. As I was doing this, it repeated back to me everything I said, saying things like "if I understand correctly you're showing me a fi…

> foo dot see see pee Well there's your problem!

The ccp extension is just another c++ flavour /i

Re: Gemini 2.0: our new AI model for the agentic era

#484

Post by nuancebydefault »

losvedir wrote: post #446
mklepaczewski wrote: post #437

Earlier quoted context omitted.

Interesting! I see how this could work for inattentive procrastinators. By "inattentive procrastinators", I mean people who are easily distracted and forget that they need to work on their tasks. Once reminded, they return to their tasks without much fuss. However, I doubt it would work for hedonistic procrastinators. When body doubling, hedonistic procrastinators rely on social pressure to be productive. Using AI li…

You don't necessarily need to believe the AI is a human for it to tickle the ingrained social instincts you're looking for. For example, I'm quite aware that AI's are just tools, and yet I still feel a strong need to be "polite" in my requests to ChatGPT. "Please do ...." or "Can you...?" and even "Thanks, that worked! Now can you..." etc.

You only do that politeness as a novice.

My questions to copilot.ms.com today are more like the following, still works like a charm...

"I have cpp code: and i get error . Is this wrong smart ponitor?"

[elaborate answer with nice examples]

"Works. "

Re: Gemini 2.0: our new AI model for the agentic era

#485

Post by SimianSci »

mark_l_watson wrote: post #345
SimianSci wrote: post #225

Earlier quoted context omitted.

I think Apple is uniquely disadvantaged in the AI race to a point people dont realize. They have less training data to use, having famously been focused on privacy for its users and thus having no particular advantage in this space due to not having customer data to train on. They have little to no cloud business, and while they operate a couple of services for their users, they do not have the infrastructure scale t…

It is likely Apple can get additional data by creating synthetic data for user interactions. About 7 years ago I trained GAN models to generate synthetic data, and it worked so well. The state of the art has increased a lot in 7 years, so Apple will be fine.

For a while there I would have been in agreeance with you, but the thought that models can be trained purely on synthetic data has shown to be wrong on multiple levels. Synthetic data needs to be reviewed by individuals to ensure data quality, significantly reducing the speed at which an organization can adopt training data. Reasonable engineers would suggest that the answer to this is to have other language models review the synthetic data, but we have seen that this is what leads to model collapse due to compounding issues around hallucinations.

At best Synthetic data is a "slow follow" for training a model due to the need for human review, but a competitive model, it does not make.

Re: Gemini 2.0: our new AI model for the agentic era

#486

Post by Jeff_Brown »

sky2224 wrote: post #397
modeless wrote: post #383

https://aistudio.google.com/live is by far the coolest thing here. You can just go there and share your screen or camera and have a running live voice conversation with Gemini about anything you're looking at. As much as you want, for free. I just tried having it teach me how to use Blender. It seems like it could actually be super helpful for beginners, as it has decent knowledge of the toolbars and keyboard shortcu…

THIS is the thing I'm excited for with AI. I'm someone that becomes about 5x more productive when I have a person watching or just checking in on me (even if they're just hovering there). Having AI to basically be that "parent" to kick me into gear would be so helpful. 90% of the time my problems are because I need someone to help keep the gears turning for me, but there isn't always someone available. This has the p…

I'm intrigued to know whether that actually ends up working. I am something like that myself, but I don't know whether it is an effect of getting feedback or of having a person behind the feedback.

Re: Gemini 2.0: our new AI model for the agentic era

#487

Post by rowanG077 »

whimsicalism wrote: post #194
rowanG077 wrote: post #166

Earlier quoted context omitted.

That makes no sense. Inference cost dwarf training cost if you have a succesfull product pretty quickly. Afaik there is no commodity hardware that can run state of the art models like chatgpt-o1.

> Afaik there is no commodity hardware that can run state of the art models like chatgpt-o1. Stack enough GPUs and any of them can run o1. Building a chip to infer LLMs is much easier than building a training chip. Just because one cost dwarfs another does not mean that this is where the most marginal value from developing a better chip will be, especially if other people are just doing it for you. Google gets a good…

Chip level is only a tiny part of the story. Training can happen with a big boy variant of "it works on my machine". Inference require a world wide network of GPUs. Chip level is the last thing you will be worrying about.

Re: Gemini 2.0: our new AI model for the agentic era

#488

Post by dwaltrip »

dtquad wrote: post #435
nolist_policy wrote: post #431

Earlier quoted context omitted.

You just rediscovered that LLMs become much stupider when using images as input. I think that has been already shown for gpt-4 as well.

Even when using the web search tool GPT-4 becomes stupider. Do they use a dumber model for tool/vision?

I’m guessing that it’s just a much harder problem. Images often contain more information but it is far less structured and refined than language.

The transformation process that occurs when people speak and write is incredibly rich and complex. Compared to images which are essentially just the outputs of cameras or screen captures — there isn’t an “intelligent” transformation process occurring.

Re: Gemini 2.0: our new AI model for the agentic era

#489

Post by AH4oFVbPT4f8 »

Brotkrumen wrote: post #388
AH4oFVbPT4f8 wrote: post #355

Earlier quoted context omitted.

Why is Google THE AI play as you put it? I don't agree or disagree, just wanting to understand your perspective.

That conversation reads like two consultant LLMs talking past each other.

I actually wanted to know, real person proof user: AH4oFVbPT4f8 created: August 12, 2013

I'm going back and forth between the different models seeing which works best for me but I'm trying to learn how to read and use other people's feedback in making their decisions.

Re: Gemini 2.0: our new AI model for the agentic era

#490

Post by kridsdale1 »

cryptozeus wrote: post #473
modeless wrote: post #383

https://aistudio.google.com/live is by far the coolest thing here. You can just go there and share your screen or camera and have a running live voice conversation with Gemini about anything you're looking at. As much as you want, for free. I just tried having it teach me how to use Blender. It seems like it could actually be super helpful for beginners, as it has decent knowledge of the toolbars and keyboard shortcu…

This comment is better than entire ad google just showed. Who is still pointing at the building with camera and asking what is this building?

I do that in Manhattan. I also do it for yonder mountains.
Post reply on HN