Gemini "duck" demo was not done in realtime or with voice • Hacker News Board

Re: Gemini "duck" demo was not done in realtime or with voice

#291

Post by robbomacrae » Thu, Dec 07, 2023, 9:47 PM UTC

AndrewKemendo wrote: post #228

mvdtnz wrote: post #76

Earlier quoted context omitted.

Showing a product in its best light is one thing. Demonstrating a mode of operation that doesn't exist is entirely another. It would be like if a demo of your swipe keyboard included telepathic mind control for correcting errors.

I’m not sure I’d agree that what they showed will never be possible and in fact my whole point is that I think Google can most likely deliver on that in this specific case. Chalk it up to my experience in the space, but from what I can see it looks like something Google can actually execute on (unlike many areas where they fail on product regularly). I would agree completely that it’s not ready for consumers the way…

The voice interaction part didn't look a far cry from what we are doing with Dynamic Interaction at SoundHound. Because of this I assumed (like many it seems) that they had caught up.

And it's dangerous to assume they can just "deliver later". It's not that simple. If it is why not bake it in right now instead of committing fraud?

This is damaging to companies that walk the walk and then people have literally said to me "but what about that Gemini"? and dismiss our work.

Re: Gemini "duck" demo was not done in realtime or with voice

#292

Post by bee_rider » Thu, Dec 07, 2023, 9:47 PM UTC

andrewmutz wrote: post #283

colonwqbang wrote: post #276

Earlier quoted context omitted.

Is an aircraft carrier made of a material that is less dense than water?

Is an aircraft carrier made of metal and air? Or just metal?

Where’s the distinction between the air that is part of the boat, and the air that is not? If the air is included in the boat, should we all be wearing life vests?

Re: Gemini "duck" demo was not done in realtime or with voice

#293

Post by bradhe » Thu, Dec 07, 2023, 9:48 PM UTC

DonnyV wrote: post #259

This is so crazy. Google invented transformers which is the bases for all these models. How do they keep fumbling like this over and over. Google Docs created in 2006! Microsoft is eating their lunch. Google creates the ability to change VM's in place and makes a fully automated datacenter. Amazon and Microsoft are killing them in the cloud. Google has been working on self driving longer than anyone. Tesla is catchin…

Microsoft eating Google's lunch on documents is laughable at best. Not to mention it confuses the entire timeline of office productivity software??

Re: Gemini "duck" demo was not done in realtime or with voice

#294

Post by anon373839 » Thu, Dec 07, 2023, 9:49 PM UTC

jmathai wrote: post #166

stefan_ wrote: post #148

Earlier quoted context omitted.

You think you ask your legal assistant to find some precedents related to your current case and they will come back with an A4 page full of made up cases that sound vaguely related and convincing but are not real ? I don't think you understand the failure case at all.

That example seems a bit hyperbolic. Do you think lawyers who leverage ChatGPT will take the made up cases and present them to a judge without doing some additional research? What I'm saying is that the tolerance for mistakes is strongly correlated to the value ChatGPT creates. I think both will need to be improved but there's probably more opportunity in creating higher value. I don't have a horse in the race.

> Do you think lawyers who leverage ChatGPT will take the made up cases and present them to a judge without doing some additional research

I really don’t recommend using ChatGPT (even GPT-4) for legal research or analysis. It’s simply terrible at it if you’re examining anything remotely novel. I suspect there is a valuable RAG application to be built for searching and summarizing case law, but the “reasoning” ability and stored knowledge of these models is worse than useless.

Re: Gemini "duck" demo was not done in realtime or with voice

#295

Post by w10-1 » Thu, Dec 07, 2023, 9:50 PM UTC

DonnyV wrote: post #259

This is so crazy. Google invented transformers which is the bases for all these models. How do they keep fumbling like this over and over. Google Docs created in 2006! Microsoft is eating their lunch. Google creates the ability to change VM's in place and makes a fully automated datacenter. Amazon and Microsoft are killing them in the cloud. Google has been working on self driving longer than anyone. Tesla is catchin…

Isn't it always easier to learn from others' mistakes?

Google has the problem that it's typically the first to encounter a problem, and it has the resources to approach it (from search), but the incentive to monetize it (to get away from depending entirely on search revenue). And, management.

Re: Gemini "duck" demo was not done in realtime or with voice

#296

Post by modeless » Thu, Dec 07, 2023, 9:50 PM UTC

dragonwriter wrote: post #271

modeless wrote: post #186

Earlier quoted context omitted.

I'm not a fan of current approaches here. "Chain of thought" or other approaches where the model does all its thinking using a literal internal monologue in text seem like a dead end. Humans do most of their thinking non-verbally and we need to figure out how to get these models to think non-verbally too. Unfortunately it seems that Gemini represents no progress in this direction.

> "Chain of thought" or other approaches where the model does all its thinking using a literal internal monologue in text seem like a dead end. Humans do most of their thinking non-verbally and we need to figure out how to get these models to think non-verbally too. Insofar as we can say that models think at all between the input and the stream of tokens output, they do it nonverbally. Forcing the structure of reduce…

You're right, the models do think non-verbally. However, crucially, they can only do so for a fixed amount of time for each output token. What's needed is a way for them to think non-verbally continuously, and decide for themselves when they've done enough thinking to output the next token.

Re: Gemini "duck" demo was not done in realtime or with voice

#297

Post by lm28469 » Thu, Dec 07, 2023, 9:50 PM UTC

bongodongobob wrote: post #205

bitvoid wrote: post #188

Earlier quoted context omitted.

> The number of non-tech ppl I've heard directly reference ChatGPT now is absolutely shocking. The problem is that a lot of those people will take ChatGPT output at face value. They are wholly unaware that of its inaccuracies or that it hallucinates. I've seen it too many times in the relatively short amount of time that ChatGPT has been around.

So what? People do this with Facebook news too. That's a people problem, not an LLM problem.

Literally everything is a "people problem"

You can kill people with a fork, it doesn't mean you should legally be allowed to own a nuclear bomb "because it's just the same". The problem always come from scale and accessibility

Re: Gemini "duck" demo was not done in realtime or with voice

#298

Post by dylan604 » Thu, Dec 07, 2023, 9:50 PM UTC

zozbot234 wrote: post #267

dylan604 wrote: post #227

Earlier quoted context omitted.

People on social media are absolutely 100% posting things deliberately to fuck with people. They are actively seeking to confuse people, cause chaos, divisiveness, and other ill intended purposes. Unless you're saying that the LLM developers are actively doing the same thing, I don't think comparing what people find on the socials vs getting back as a response from a chatBot is a logical comparison at all

How is that any different from what these AI chatbots are doing? They make stuff up that they predict will be rewarded highly by humans who look at it. This is exactly what leads to truisms like "rubber duckies are made of a material that floats over water" - which looks like it should be correct, even though it's wrong. It really is no different from Facebook memes that are devised to get a rise out of people and be…

Because we shouldn't be striving to make mediocrity. We should be striving to build better. Unless the devs of the bots are wanting to have a bot built on trying to deceive people, I just don't see the purpose of this. If we can "train" a bot and fine tune it, we should be fine tuning truth and telling it what absolutely is bullshit.

To avoid the darker topics to keep the conversation on the rails, if there were a misinformation campaign that was trying to state that the Earth's sky is red, then the fine tuning should be able to inform that this is clearly fake so when quoting this it should be stated as incorrect information that is out there. This kind of development should be how we can clean up the fake, but nope, we're seemingly quite happy at accepting it. At least that's how your question comes off to me.

Re: Gemini "duck" demo was not done in realtime or with voice

#299

Post by recursive » Thu, Dec 07, 2023, 9:51 PM UTC

rowanG077 wrote: post #226

modeless wrote: post #94

That's not the only thing wrong. Gemini makes a false statement in the video, serving as a great demonstration of how these models still outright lie so frequently, so casually, and so convincingly that you won't notice, even if you have a whole team of researchers and video editors reviewing the output. It's the single biggest problem with LLMs and Gemini isn't solving it. You simply can't rely on them when correctn…

The duck is indeed made of a material that is less dense. Namely water and air. If you go to such technical routes your definition is wrong too. It doesn't float because it contains air. If you poke in the head of the duck it will sink. Even though at all times it contains air.

The duck is made of water and air? Which duck are we talking about here.

Re: Gemini "duck" demo was not done in realtime or with voice

#300

Post by robbomacrae » Thu, Dec 07, 2023, 9:52 PM UTC

kozikow wrote: post #204

htk wrote: post #86

The whole Gemini webpage and contents felt weird to me, it's in the uncanny valley of trying to look and feel like an Apple marketing piece. The hyperbolic language, surgically precise ethnic/gender diversity, unnecessary animations and the sales pitch from the CEO felt like a small player in the field trying to pass as a big one.

It's funny because now the OpenAI keynote feels like it's emulating the Google keynotes from 5 years ago. Google Keynote feels like it's emulating the Apple keynote from 5 years ago. And the Apple keynote looks like robots just out of an uncanny valley pretending to be humans - just like keynotes might look in 5 years, but actually made by AI. Apple is always ahead of the curve in keynote trends.

The more I think about this the more it rings true...