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

Google is winning on every AI front

thealgorithmicbridge.com

Post reply on HN
Print view View original

641–650 of 851 posts

Re: Google is winning on every AI front

#641

Post by sva_ »

whiplash451 wrote: post #640
sva_ wrote: post #638

Earlier quoted context omitted.

Would be illegal in Germany ('Schleichwerbung') and perhaps the EU? I think it is actually covered in EU AI act article 5 (a): > [...] an AI system that deploys subliminal techniques beyond a person’s consciousness or purposefully manipulative or deceptive techniques, with the objective, or the effect of materially distorting the behaviour of a person or a group of persons by appreciably impairing their ability to ma…

The trick is in the « materially ». The inability to demonstrate incrementality in advertising is going to come in very handy to dodge this rule.

Hmm yeah I guess I wasn't completely aware of that term and its implications. That seems like a pretty weird qualifier for such a law. Now it kind of makes it sound like the law wants to prevent people using AI in a way that makes your grandma transfer her life savings to them.

Clearly, most LLMs would work in small increments with compounding effects.

Re: Google is winning on every AI front

#642

Post by sva_ »

h2zizzle wrote: post #627
sva_ wrote: post #610

It is sort of funny to me how the sentiment about whoever seems to be leading in ML changes so frequently (in particular here on HN.) A couple months ago it felt like people were sure that Google completely fucked it up for themselves (especially due to the fact that they invented the transformer but didn't productize it themselves at first.) For a short while, Claude was the best thing since sliced cheese, then Deep…

You could also be seeing waves of various astroturf campaigns.

Yeah... I wish there were laws that would require disclosure of such behavior. Might be tricky to implement though, and probably contradicts the interests of politicians.

Re: Google is winning on every AI front

#643

Post by mota7 »

jxjnskkzxxhx wrote: post #374
thunderbird120 wrote: post #10

This article doesn't mention TPUs anywhere. I don't think it's obvious for people outside of google's ecosystem just how extraordinarily good the JAX + TPU ecosystem is. Google several structural advantages over other major players, but the largest one is that they roll their own compute solution which is actually very mature and competitive. TPUs are extremely good at both training and inference[1] especially at sca…

I've used Jax quite a bit and it's so much better than tf/pytorch. Now for the life of me, I still haven't been able to understan what a TPU is. Is it Google's marketing term for a GPU? Or is it something different entirely?

There's basically a difference in philosophy. GPU chips have a bunch of cores, each of which is semi-capable, whereas TPU chips have (effectively) one enormous core.

So GPUs have ~120 small systolic arrays, one per SM (aka, a tensorcore), plus passable off-chip bandwidth (aka 16 lines of PCI).

Where has TPUs have one honking big systolic array, plus large amounts of off-chip bandwidth.

This roughly translates to GPUs being better if you're doing a bunch of different small-ish things in parallel, but TPUs are better if you're doing lots of large matrix multiplies.

Re: Google is winning on every AI front

#644

Post by gundmc »

OJFord wrote: post #210
AnonymousPlanet wrote: post #106

Earlier quoted context omitted.

GP said they were in Sydney. As far as I know, the week officially starts on Monday in Australia. That's also the case for most of Europe, BTW. Maybe their locale wasn't set right or it's another case of American software assuming weird American standards for the rest of the world.

I didn't know this about the USA - but it's still called 'the weekend'? (GP uses it at least) ..Even though 50% of it is apparently 'the weekstart'?

Yes, the weekend is Saturday and Sunday in the United States. I guess you could consider it like "bookends", for us it's the start and end of the week.

Re: Google is winning on every AI front

#645

Post by ramraj07 »

chasd00 wrote: post #218
ramraj07 wrote: post #195

Earlier quoted context omitted.

Time for my next round of Evals then. I had a 40 PR coding streak last weekend with mostly o3-mini-pro, will test the latest 2.5 now.

PR = pull request? So every bit of garbage from the LLM, over and over, resulted in an individual pull request? Why not just do one when your branch is finally right?

A pull request in my workplace is an actual feature/enhancement/bug-fix. That many PRs means I shipped that many features or enhancements.

I suppose you don't know what a PR is because you likely still work in an environment without modern version control, probably just now migrating your rants from vim vs emacs to crapping on vibe coding.

In my experience, AI today is an intelligence multiplier. A lot of folks just need to look back at the zero they keep multiplying and getting zero back to understand why they don't get the hype.

Re: Google is winning on every AI front

#646

Post by famouswaffles »

matthewdgreen wrote: post #621
famouswaffles wrote: post #526

Earlier quoted context omitted.

Mostly only social media/messaging sites report daily active users regularly. Everything else usually reports monthly active users at best. >in the last day? If you’re only using something once per week, it probably isn’t that important to you. No, something I use on a weekly basis (which is not necessarily just once a week) is pretty important to me and spinning it otherwise is bizarre. Google is the frontend to the…

I am entirely confused by this. ChatGPT is absolutely unimportant to me. I don't use it for any serious work, I don't use it for search, I find its output to still be mostly a novelty. Even coding questions I mostly solve using StackExchange searches because I've been burned using it a couple of times in esoteric areas. In the few areas where I actually did want some solid LLM output, I used Claude. If ChatGPT disapp…

>I am entirely confused by this. ChatGPT is absolutely unimportant to me. I don't use it for any serious work, I don't use it for search, I find its output to still be mostly a novelty. Even coding questions I mostly solve using StackExchange searches because I've been burned using it a couple of times in esoteric areas. In the few areas where I actually did want some solid LLM output, I used Claude. If ChatGPT disappeared off the Internet tomorrow, I would suffer not at all.

OK? That's fine. I don't think I ever claimed you were a WAU

>And yet I probably duck into ChatGPT at least once a month or more (I see a bunch of trivial uses in 2024) mostly as a novelty.

So you are not a weekly active user then. Maybe not even a monthly active one.

>Last week I used it a bunch because my wife wanted a logo for a new website. But I could have easily made that logo with another service.

Maybe[1], but you didn't. And I doubt your wife needs a new logo every week so again not a weekly active user.

>ChatGPT serves the same role to me as dozens of other replaceable Internet services that I probably duck into on a weekly basis (e.g., random finance websites, meme generators)but have no essential need for whatsoever.

You visit the same exact meme generator or finance site every week? If so, then that site is pretty important to you. If not, then again you're not a weekly active user to it.

If you visit a (but not the same) meme generator every week then clearly creating memes is important to you because I've never visited one in my life.

>And if I did have an essential need for it, there are at least four well-funded competitors with all the same capabilities, and modestly weaker open weight models.

There are well funded alternatives to Google Search too but how many use anything else? Rarely does any valuable niche have no competition.

>It is really your view that "any service you use at least once a week must be really important to you?" I bet if you sat down and looked at your web history, you'd find dozens that aren't.

Yeah it is and so far, you've not actually said anything to indicate the contrary.

[1]ChatGPT had an image generation update recently that made it capable of doing things other services can't. Good chance you could not in fact do what you did (to the same satisfaction) elsewhere. But that's beside my point.

Re: Google is winning on every AI front

#647

Post by ramraj07 »

scrollop wrote: post #304
ramraj07 wrote: post #159

Earlier quoted context omitted.

I don't know what you did there, but clearly being ex OpenAI isn't the intellectual or product flex it is: I and every other smart person I know still use ChatGPT (paid) because even now it's the best at what it does and we keep trying Google and Claude and keep coming back. They got and as of now continue to get things right for the most part. If you still aren'ĥt seeing it maybe you should introspect what you're mi…

I use a service where users can choose any frontier model, and OpenAI models haven't been the most used model for over half a year - it was sonnet until gemini 2.5 pro came out, recently. Not sure whether you have your perspective because you're invested os much into OpenAI, however the general consensus is that gemini 2.5 pro is the top model at the moment, including all the AI reviews and OpenAI is barely mentioned…

Are you able to use o3-mini-high through these tools?

Re: Google is winning on every AI front

#648

Post by joenot443 »

h2zizzle wrote: post #627
sva_ wrote: post #610

It is sort of funny to me how the sentiment about whoever seems to be leading in ML changes so frequently (in particular here on HN.) A couple months ago it felt like people were sure that Google completely fucked it up for themselves (especially due to the fact that they invented the transformer but didn't productize it themselves at first.) For a short while, Claude was the best thing since sliced cheese, then Deep…

You could also be seeing waves of various astroturf campaigns.

Personally, I don't really think there's a team at Google, nor at OpenAI, paying for "astroturfing" on sites like HN.

What are the rough steps through which you see this working? I see people talking about "astroturfing" all the time without much explanation on the mechanisms. So roughly, there are employees paid solely to post on social media like HN trying to push the needle in one direction or another?

Re: Google is winning on every AI front

#649

Post by nothercastle »

coredog64 wrote: post #609
nothercastle wrote: post #441

Earlier quoted context omitted.

I agree the market for 10% better AI isn’t that great but the cost to get there is. An 80% as good model at 10% or even 5% the cost will win every time in the current environment. Most businesses don’t even have a clear use case for AI they just use it because the competition is and there is a FOMO effect

> Most businesses don’t even have a clear use case for AI they just use it because the competition is and there is a FOMO effect I consult in this space and 80-90% of what I see is chat bots and RAG.

That’s exactly what I’d expect. Honestly Ai chat bots seems unnecessarily risky because you never really know what they might say on your behalf.

Re: Google is winning on every AI front

#650

Post by int_19h »

famouswaffles wrote: post #600
codelord wrote: post #31

As an Ex-OpenAI employee I agree with this. Most of the top ML talent at OpenAI already have left to either do their own thing or join other startups. A few are still there but I doubt if they'll be around in a year. The main successful product from OpenAI is the ChatGPT app, but there's a limit on how much you can charge people for subscription fees. I think soon people expect this service to be provided for free an…

Open AI don't always have the best models (especially for programming) but they've consistently had the best product/user experience. And even in the model front, other companies seem to play catchup more than anything most of the time.

The best user experience for what?

The most practical use case for generative AI today is coding assistants, and if you look at that market, the best offerings are third-party IDEs that build on top of models they don't own. E.g. Cursor + Gemini 2.5.

On the model front, it used to be the case that other companies were playing catch-up with OpenAI. I was one of the people consistently pointing out that "better than GPT o1" on a bunch of benchmarks does not reliably translate to actual improvements when you try to use them. But this is no longer the case, either - Gemini 2.5 is really that good, and Claude is also beating them in some real world scenarios.

Post reply on HN