Google is winning on every AI front

This article doesn&#x27;t mention TPUs anywhere. I don&#x27;t think it&#x27;s obvious for people outside of google&#x27;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 scale. Google&#x27;s ability to tailor their mature hardware to exactly what they need gives them a massive leg up on competition. AI companies fundamentally have to answer the question &quot;what can you do that no one else can?&quot;. Google&#x27;s hardware advantage provides an actual answer to that question which can&#x27;t be erased the next time someone drops a new model onto huggingface.<p>[1]<a href="https:&#x2F;&#x2F;blog.google&#x2F;products&#x2F;google-cloud&#x2F;ironwood-tpu-age-of-inference&#x2F;" rel="nofollow">https:&#x2F;&#x2F;blog.google&#x2F;products&#x2F;google-cloud&#x2F;ironwood-tpu-age-o...</a>

Google is winning on every front except... marketing (Google has a chatbot?), trust (who knew the founding fathers were so diverse?), safety (where&#x27;s the 2.5 Pro model card?), market share (fully one in ten internet users on the planet are weekly ChatGPT users), and, well, vibes (who&#x27;s rooting for big G, exactly?).<p>But I will admit, Gemini Pro 2.5 is a legit good model. So, hats off for that.

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&#x27;ll be around in a year. The main successful product from OpenAI is the ChatGPT app, but there&#x27;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 and ads would become the main option to make money out of chatbots. The whole time that I was at OpenAI until now GOOG has been the only individual stock that I&#x27;ve been holding. Despite the threat to their search business I think they&#x27;ll bounce back because they have a lot of cards to play. OpenAI is an annoyance for Google, because they are willing to burn money to get users. Google can&#x27;t as easily burn money, since they already have billions of users, but also they are a public company and have to answer to investors. But I doubt if OpenAI investors would sign up to give more money to be burned in a year. Google just needs to ease off on the red tape and make their innovations available to users as fast as they can. (And don&#x27;t let me get started with Sam Altman.)

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&#x27;t productize it themselves at first.)<p>For a short while, Claude was the best thing since sliced cheese, then Deepseek was the shit, and now seemingly OpenAI really falls out of favor. It kinda feels to me like people cast their judgement too early (perhaps again in this case.) I guess these are the hypecycles...<p>Google is killing it right now, I agree. But the world might appear completely different in three months.

Several people have suggested that LLMs might end up ad-supported. I&#x27;ll point out that &quot;ad supported&quot; might be <i>incredibly</i> subtle&#x2F;insidious when applied to LLMs:<p>An LLM-based &quot;adsense&quot; could:<p><pre><code> 1. Maintain a list of sponsors looking to buy ads 2. Maintain a profile of users&#x2F;ad targets 3. Monitor all inputs&#x2F;outputs 4. Insert &quot;recommendations&quot; (ads) smoothly&#x2F;imperceptibly in the course of normal conversation </code></pre> No one would ever need to&#x2F;be able to know if the output:<p>&quot;In order to increase hip flexibility, you might consider taking up yoga.&quot;<p>Was generated because it might lead to the question:<p>&quot;What kind of yoga equipment could I use for that?&quot;<p>Which could then lead to the output:<p>&quot;You might want to get a yoga mat and foam blocks. I can describe some of the best moves for hips, or make some recommendations for foam blocks you need to do those moves?&quot;<p>The above is ham-handed compared to what an LLM could do.

I was a loyal Claude user until I decided to try Gemini 2.5. &quot;After all&quot;, I thought, &quot;I already use a Pixel phone, so it&#x27;s integrated with Android. And with Google Drive. And I can get it through my Google One subscription.&quot;<p>And now that I&#x27;m on it, I don&#x27;t think I&#x27;m going back. Google did it again.

&gt; Gemini 2.5 Pro in Deep Research mode is twice as good as OpenAI’s Deep Research<p>That matches my impression. For the past month or two, I have been running informal side-by-side tests of the Deep Research products from OpenAI, Perplexity, and Google. OpenAI was clearly winning—more complete and incisive, and no hallucinated sources that I noticed.<p>That changed a few days ago, when Google switched their Deep Research over to Gemini 2.5 Pro Experimental. While OpenAI’s and Perplexity’s reports are still pretty good, Google’s usually seem deeper, more complete, and more incisive.<p>My prompting technique, by the way, is to first explain to a regular model the problem I’m interested in and ask it to write a full prompt that can be given to a reasoning LLM that can search the web. I check the suggested prompt, make a change or two, and then feed it to the Deep Research models.<p>One thing I’ve been playing with is asking for reports that discuss and connect three disparate topics. Below are the reports that the three Deep Research models gave me just now on surrealism, Freudian dream theory, and AI image prompt engineering. Deciding which is best is left as an exercise to the reader.<p>OpenAI:<p><a href="https:&#x2F;&#x2F;chatgpt.com&#x2F;share&#x2F;67fa21eb-18a4-8011-9a97-9f8b051ad3d3" rel="nofollow">https:&#x2F;&#x2F;chatgpt.com&#x2F;share&#x2F;67fa21eb-18a4-8011-9a97-9f8b051ad3...</a><p>Google:<p><a href="https:&#x2F;&#x2F;docs.google.com&#x2F;document&#x2F;d&#x2F;10mF_qThVcoJ5ouPMW-xKg7Cy0hKtnLo73QNikw8H69s&#x2F;edit?usp=sharing" rel="nofollow">https:&#x2F;&#x2F;docs.google.com&#x2F;document&#x2F;d&#x2F;10mF_qThVcoJ5ouPMW-xKg7Cy...</a><p>Perplexity:<p><a href="https:&#x2F;&#x2F;www.perplexity.ai&#x2F;search&#x2F;subject-analytical-report-inte-3v3iGskiQAKqpbxsg6nQxA#0" rel="nofollow">https:&#x2F;&#x2F;www.perplexity.ai&#x2F;search&#x2F;subject-analytical-report-i...</a>

Apart from Gemini 2.5 Pro they have a decent Jack-of-all-trades master of none&#x2F;price Gemini 2.0 Flash.<p>1) is dirty cheap ($0.1M&#x2F;$0.4M),<p>2) is multimodal (image and audio),<p>3) reliable rate limit (comparing to OSS ml ai providers),<p>4) fast (200 tokens&#x2F;s).<p>5) if need realtime API they provide as well for more expensive price (audio-to-audio)<p>It&#x27;s my go to model for using as an API for some apps&#x2F;products. <a href="https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;models&#x2F;gemini-2-0-flash&#x2F;providers" rel="nofollow">https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;models&#x2F;gemini-2-0-flash&#x2F;provid...</a>

Gemini 2.5 pro is as powerful as everybody says. I still also use Claude Sonnet 3.7 only because the Gemini web UI has issues... (Imagine creating the best AI and then not allowing to attach Python or C files if not renamed .txt) but the way the model is better than anyone else is a &quot;that&#x27;s another league&quot; experience. They have the biggest search engine and YouTube to leverage the power of the AI they are developing. At this point I believe too that they are likely to win the race.

I&#x27;m trying Imagen 3 to add pictures to a presentation in Google Slides, and it&#x27;s making such basic mistakes that I thought image models weren&#x27;t making any more by now. I tried for half an hour to prompt it into generating an illustration of a Thinkpad facing with the back to the viewer, so the keyboard is not visible. It couldn&#x27;t do it, it would always make the keyboard face towards the viewer. Or you ask for an illustration of an animal pointing a finger, and it gives it an additional arm. Meanwhile you ask OpenAI to ghiblify a picture while changing the setting and adding 5 other things, and it absolutely nails it.

Google will need a far better LLM than OpenAI to throw them decisively off the AI throne, just like another company would need a far better search engine than Google to throw them off the search throne. ChatGPT is now the 7th highest ranking website on the planet - does anyone outside the HN crowd know about Google AI Studio?<p>Brands matter, and when regular people think AI, they think of OpenAI before they think Google, even if Google has more AI talents and scores better on tests.<p>And isn&#x27;t it good? Who wants a world where the same handful of companies dominate all tech?

As long as Google continues to hamstring themselves with censorship for no reason, I can&#x27;t use their products. The other day I asked gemini 2.5 pro &quot;which british ceo said that his company&#x27;s products were bad&quot; and the response was<p>&gt; I&#x27;m just a language model, so I can&#x27;t help you with that.<p><a href="https:&#x2F;&#x2F;g.co&#x2F;gemini&#x2F;share&#x2F;cb3afc3e7f78" rel="nofollow">https:&#x2F;&#x2F;g.co&#x2F;gemini&#x2F;share&#x2F;cb3afc3e7f78</a><p>Chatgpt 4o correctly identified the guy as Ratner and provided the relevant quotes.

In my experience Claude 3.7 is far superior for coding than Gemini 2.5. I tried it in Cursor and I wanted it to work, as a recent ex-Googler. I repeatedly found it inferior. I think it’s still behind Claud 3.5 for coding.<p>It would decide arbitrarily not to finish tasks and suggest that I do them. It made simple errors and failed to catch them.

I haven&#x27;t seen anyone else integrate LLMs properly with email and calendar other than Google.<p>I had an interesting experience: I was taking out a loan for some solar panels and there were some complicated instructions across multiple emails. I asked Gemini to summarise exactly what I had to do. It looked through my emails and told me I had to go to the web site for local rebate scheme and apply there. It even emphasised that it was important that I do that. I scoffed at it because I thought my installer was going to do that and wrote it off. A few weeks later, guess what: the installer calls me because they can&#x27;t see the rebate application in their portal and want me to go check that I applied for it (!). Sure enough, I missed the language in the email telling me to do that and had to do exactly what Gemini had said weeks ago.<p>I do think Google has a real shot here <i>because</i> they have such an integrated email and calendaring solution where everyone already assumes it&#x27;s online, fully indexed etc.

I run every query I do through all the major models, up to 10 of them at this point.<p>Benchmarks aside Gemini 2.5 Pro is a great model and now often produces the best code for me but it&#x27;s not notably better than any of the other frontier models in my testing each of which tends to have their own strengths and weaknesses.<p>And Google&#x27;s wrapper around Gemini is easily the most frustrating of any of the major AI companies. It&#x27;s content guardrails are annoying and I just learned yesterday it won&#x27;t let you upload json files for whatever reason (change the extension to txt without modifying the contents in any way and it works just fine).

I feel the article presents the data selectively in some places. Two examples:<p>* The article compares Gemini 2.5 Pro Experimental to DeepSeek-R1 in accuracy benchmarks. Then, when the comparison becomes about cost, it compares Gemini 2.0 Flash to DeepSeek-R1.<p>* In throughput numbers, DeepSeek-R1 is quoted at 24 tok&#x2F;s. There are half a dozen providers, who give you easily 100+ tok&#x2F;s and at scale.<p>There&#x27;s no doubt that Gemini 2.5 Pro Experimental is a state of the art model. I just think it&#x27;s very hard to win on <i>every</i> AI front these days.

No it&#x27;s not obvious at all Google is winning AI on every front. There is few stuff Google is systemically behind: 1) UX 2) product and use case innovation<p>I just open Google Gemini Android app and asked to generate a JS script with Gemini 2 Flash and did the same with ChatGPT.<p>Gemini did not highlighted with colors the code. ChatGPT did highlighted with colors the code.<p>Colors in code are extremely useful to grok the code and have a nice DX.<p>I&#x27;m sure if I dig into Gemini&#x27;s product I&#x27;ll find dozens of UX&#x2F;DX ways in which ChatGPT is better.<p>Google is still playing catch-up with LLM products. ChatGPT is still the one making the announcements and then Gemini doing the same UX&#x2F;use case enhancements weeks&#x2F;months later.

&gt; I felt Demis Hassabis was trustworthy in a way Sam Altman couldn&#x27;t be—a true scientist, not a businessman<p>Not that I think Demis is or is not trustworthy, but I think it’s a bit foolish to believe it would be allowed to matter.

I only AI for one reason since I&#x27;m retired and live alone: Life-like chats with a reasonable approximation of a knowledgeable friend. With the new memory features ChatGPT excels at that. I&#x27;m not even sure Google cares about that; that goes to show how little of it I&#x27;ve noticed with Google.

May be it&#x27;s my luck but I found a glaring issue with Gemini 2.5 Pro in AI Studio.<p>I asked it whether a language feature in Zig was available. It answered yes and proceeded to generate a whole working sample code. I compiled it and got an error. Reported the error and it said the error showed I typed it wrong and asked me to make sure it&#x27;s typed correctly. Eh?! It&#x27;s a copy-and-paste. I confirmed again it&#x27;s wrong. It then said it must be my compiler version was too old. Nope, using the latest. It then said very convincingly that based on its extensively research into the language official documentation, official examples, and release notes, the feature must exist. I asked it to show me the reference materials it used to draw the conclusion. None of links it gave were valid. I told it they were wrong. It gave back another set of links and claimed it had checked the links to make sure they are alive. The links were alive but didn&#x27;t contain any mention of the feature. I let it know again. It admitted couldn&#x27;t find the mentioned feature. But it insisted the feature had been merged in a PR. The PR link it gave was unrelated. I let it know. It gave me another 3 PR&#x27;s and said one mentioned something related so the feature must be in. At the point I gave up.<p>The issue was that it sounded very convincing and stated &quot;facts&quot; very confidently, with backings to documents and other resources even if they were wrong or irrelevant. Even when told it gave the wrong info, it would double down and made up some BS reference material to back up its claim.

I think my experience has been different from everyone else. As an owner of a pixel phone and multiple Google accounts, I wanted this to be true. But Gemini has been super inconsistent with tasks that are trivial for Google Assistant. I even bought the $26 AI plan for my account to help with some proofreading and it&#x27;s been awful compared to ChatGPT. I&#x27;m about to cancel it.

The author mentioned AlphaGo and Alpha Zero without mentioning OpenAI gym and OpenAI Five.<p>Those products show OpenAI was innovating and leading in RL at that stage around 2017 to 2019.<p><a href="https:&#x2F;&#x2F;github.com&#x2F;openai&#x2F;gym">https:&#x2F;&#x2F;github.com&#x2F;openai&#x2F;gym</a><p><a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;OpenAI_Five" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;OpenAI_Five</a>

&gt; Add to the above that Gemini 2.5, compared to models of its category, is fast and cheap—I mean, they&#x27;re giving away free access!<p>A large player with massive existing streams giving away a product in a new market to undercut new entrants? Looks an awful lot like abuse of monopoly position...

I recently had to check some legal thing - I gave the pdf with law to both - chatgpt and Gemini, and I was able to convince the Gemini that my interpretation is right, but chatgpt was constantly opposing me. Later I checked and found out that my interpretation was wrong, so I&#x27;d say that chatgpt was better and moreover it spared me some problems with &quot;Polish IRS&quot;

Gemini Pro 2.5 is fantastic. I&#x27;m anti Google and a long time ChatGPT user. I use it for text review and research and it&#x27;s well ahead the competition. Let&#x27;s see how long they last giving it for free.

Not on cars, not in robotics, not in commercially deployed AI, not in enterprise investments in their cloud business.<p>They&#x27;ve got immense potential, sure. But to say that they&#x27;re winning is a bit far from reality. Right now, their Cloud AI offerings to the enterprise are technologically superior to anything else out there from AWS, but guess what? AWS seems to have significantly more %age sales growth in this space with their larger base compared to GCP with their smaller market share.<p>The same can be said across turn based chat and physical AI. OpenAI continues to be the growth leader in the consumer space and a collection of Claude + self hosted + Gemini now in the enterprise &#x2F; API space.<p>They need to be measuring themselves on moving the needle in adoption now. I&#x27;d hate for such amazing progress to stall out in a niche.

I think the key is that Google is the gateway to the internet for the entire world.<p>Think about it. Whatever you’re trying to do online, either Search, Chrome or Android are in the critical path of like 90%+ of people if not more.<p>Once AI is deeply baked into these products, which are more like the “operating system” of the internet, the race is basically over.<p>Not to mention that Google is already sitting on the largest money printer in history and they can do this all day.

The article doesn&#x27;t mention one of the most complex benchmarks - ARC challenge. All models suck in it <a href="https:&#x2F;&#x2F;arcprize.org&#x2F;leaderboard" rel="nofollow">https:&#x2F;&#x2F;arcprize.org&#x2F;leaderboard</a><p>but Gemini and Claude still suck much worse then ChatGPT models

Google is winning because LLMs without a (good) search backend are mostly useless.<p>So many LLM workloads require high quality search results (backed by efficient, relevant, complete and up-to-date indexes), and that’s Google’s muscle.

It&#x27;s always seemed to me that AI is going to be a commodity business - it&#x27;s pretty clear that any company with enough money can compete, and it seems that current LLM-based AI is levelling off in terms of capability, with the new focus being on building layers of services on top of that (e.g. deep research agents).<p>In a commodity business cost is key, and Google with their N&#x27;th generation home grown TPUs and AI-optimized datacenters have a big advantage over anyone paying NVIDIA markups for accelerators or without this level of vertical integration.

I just wish they would stop trying to use their models to help keep their terrible cloud business alive.<p>Currently my teams building 2-3 systems based on Gemini, but trying to walk a client through setting up a gcp account and provision the model for video is a horrible experience. Chat et al would break their own backs trying to give you an api key fast enough, not google. Here’s a comically bad process with several layers of permissions that nobody asked for.<p>The irony of using ChatGPT to walk through setting up Gemini for a client was a highlight for me this week.

Google is the primary target for current US anti-big-tech sentiments that are getting political traction with Lina Khan and Steve Bannon teaming up at a recent conference against US Big Tech companies. J.D. Vance has also expressed that he agrees with Lina Khan and Steve Bannon and would like to see US Big Tech companies like Google be forcibly split up.<p>What will happen with Google&#x27;s AI wing when Google inevitably gets split up in the next 4-8 years?

Gemini is really good, dev is happy, cost is low, but the mass just don&#x27;t use it, The UX feels...weird(in my opinion), and Google marketing just sucks. Look at the craze of GPT-4o image gen launch, Pretty sure none of that will ever happened with Gemini(which tbf, have extremely good imagen too!) They have weird too-careful-and-afraid-to-step-on-someone-toes approach to marketing, and that shows.<p>OpenAI have Sam as &#x27;presenter&#x27; that people can hear him talk enthusiastically, Grok have Elon, even though that not really help much. Gemini? just AI product from faceless mega corporate. People don&#x27;t feel relate to that. The emotion is part of equation of why people use something, not just function.<p>Also doesn&#x27;t help with ChatGPT name already cemented in AI history. Every Bob and Jane heard about ChatGPT even they&#x27;re use it once, but Gemini...who?

What about Grok and Chinese counterparts?<p>X data is private now which would give advantage it real-time scenarios. And Chinese have made it state-level priority.

The Rubik&#x27;s cube example is just reversing the moves.<p>Writing a visualiser and basic scrambler isn&#x27;t hard to stumble upon, there&#x27;s endless training material and not much to screw up. Writing a working solver even if you train it on examples would be hard.<p>Very funny.

Back when all the articles talked about how OpenAI swiped Google&#x27;s crown while Google sat on transformers and never productized them, I saw this future coming. Google had back then, and still has, the best research on this topic, and ultimately that was going to win the day.<p>Sure, hindsight is 20&#x2F;20, and who knows if any of these products will be big money makers vs commodities, and they may still fail at the productization of these things. Sure.<p>But insofar as productization follows great technology, Google was always going to have the upper hand here. It took many years but they did finally start coming out ahead

Google is also the only company that has had their own AI hardware that&#x27;s worked (TPU). This could lead to more cost-effective training + inference and hence better AI.

I chat with Gemini Pro 2.5 Exp in Continue.dev IntelliJ plugin. I told it to &quot;implement this method&quot; and it even suggested improvements in other files (that were included in context). I feel like talking to a person.<p>BUT more often than not, it stopped halfway (the code, so it&#x27;s unusable). I&#x27;m not sure if it&#x27;s the plugin that cannot handle the response, but it never happens with Claude.

Industrial&#x2F;commercial adoption of LLMs is quite varied and critically depends on the quality vs criticality match.<p>In healthcare, engineering, construction, manufacturing, or aviation industires adoption is mostly on the admin side for low priority&#x2F;busy work - virtually no doctors, pharmacists, nurses, engineers, technicians or even sales people use LLMs on the job. LLMs products have serious quality issues and are decades behind industrial databases, simulation and diagnostic tools.<p>On the other hand in academics, consulting, publishing, journalism, marketing, advertising, law and insurance industries it&#x27;s wildly adopted and is surging. For example, Westlaw&#x27;s Co-counsel is better at drafting and case law briefing than all but the most experienced litigators. Not only it has not replaced lawyers, but is causing a boom in hiring since the time and cost of training new associates is drastically reduced and allows firms to take on more clients with more sophisticated case work.

This reads like sports commentary.

Wow, reading these comments it seems like Gemini 2.5 Pro Exp (assuming from gemini.google.com and not from Google Ai studio) is actually worth giving a shot! Is it really that impressive of a model now?<p>I&#x27;ve been using Qwen Chat a lot for the last couple of months because I got tired of Claudes small quota for free users, ChatGPTs inferior models and absurd pricing and Geminis (the previous models) heavy guardrails and censorship, like to the point that sincerely prompts actually triggers refusal.<p>I&#x27;ll try Gemini 2.5 Pro Exp again and see how well it performs this time.<p>Also, did anyone notice that the ui of Google ai studio has changed? Can&#x27;t find any mentions of this update in the release notes <a href="https:&#x2F;&#x2F;ai.google.dev&#x2F;gemini-api&#x2F;docs&#x2F;changelog" rel="nofollow">https:&#x2F;&#x2F;ai.google.dev&#x2F;gemini-api&#x2F;docs&#x2F;changelog</a>

AI is essentially a hardware&#x2F;electricity arms race.<p>Whatever model is at the top can be surpassed if a competitor has enough compute scale. We are rapidly approaching the era where it’s difficult to have enough power in one campus. Distributed sites are needed if models continue to scale at 4.7x&#x2F;year (see Epoch.ai) simply from a power perspective. You have to put the data centers where the power is and connect them together.<p>I believe the era of distributed training is already here however not everyone will be able to distribute training to multiple sites using their scale up networks. Their scale out networks will not be ready. So it could be that we see models plateau until distributed training infra is available.<p>I see the infrastructure side of AI and based on HW build out; Google has been slow to build and is behind everywhere.

&gt; is fast and cheap—I mean, they&#x27;re giving away free access!—has a gigantic context window of 1 million tokens (only recently surpassed by Meta’s Llama 4) and it’s connected to the entire Google suite of products<p>Is this a feature? I feel like using Google&#x27;s LLM products only serves to feed their Ad machine to sell me more ads. Every cloud based office suite offers AI functionality now. Unless I&#x27;m doing something really complex&#x2F;dramatic I&#x27;m going to choose the LLM that isn&#x27;t tied to a giant machine selling me ads over the one that does every time. Chat LLM products have pretty much effectively allowed me to divorce myself from the Google Ad Machine, now that I&#x27;m free I&#x27;m not walking back willingly.

The only thing I think ChatGPT is better is native image generation, their model is able to work in the joint-space much better than any other model I have seen, but I&#x27;m sure Google will try to catch up rapidly.

Their technical progress is indeed impressive. And their price dumping of 2.5 Pro for free will have moved a lot of technical users.<p>The key question is if the can stop the decline in search or pivot their revenue streams to Gemini.

I do agree with a lot of what is said here. There are however a few things that I think will hinder Google on the long-run:<p>- The last time I checked (3-4 months ago) Gemini embedding models are probably the least reliable &#x2F; contextually aware out there - A significant chunk of the market will want the ability to use locally hosted models &#x2F; manage their own which Google currently has no play for - API documentation. Across the big managed models they are likely the least well documented model. - Allowing for more system vs. user prompts

Totally agree Gemini Pro 2.5 is currently ahead but it isn&#x27;t by a gigantic amount.<p>It&#x27;s also still uncertain whether Google can turn Gemini into a successful product that either consumers or businesses want to use. They are famously bad at translating their technological advantage into good products - for example the way they shoehorn AI chat into search just makes both worse (imo).<p>I think OpenAI has the consumers and that&#x27;ll make it easier to get business. Once they start eating into Google&#x27;s lunch with AI booked flights and hotels...

&gt; Neither OpenAI nor Anthropic have a chance at this point<p>I&#x27;m so done with articles that don&#x27;t even try to talk about probability sensibly.<p>The article doesn&#x27;t make a good case even for a watered-down version of the claim. Where is the logic?<p>Until the author puts forth his model of change and how&#x2F;why Google is unassailably ahead, I&#x27;m not buying his hyperbole.<p>&gt; When I put the Google + DeepMind picture together, I can only wonder why people, myself included, ever became so bullish on OpenAI or Anthropic or even Meta.<p>Yikes. Hindsight bias in full display.

I have been using all AI tools in my job working for ai startup and none of them are google products. May be I should give it a try, their product positioning is horrible

If Google is able to ignore the pressure to bring in revenue from AI and is able to outcompete the others at automating AI research itself, I think they will win the war. It seems that they certainly have an advantage, with limited pressure from outside investors, their own hardware stack, a constant flow of cash through the other lines of business, and a head start against most of the other giant tech companies.

Just a fun I had with Gemini 2.5 pro. I asked him to create a code for a toolchain that uses some ai at a point and it stubbornly used OpenAI api and refused to generate code using Gemini apis instead. Also it said it is model trained by google. Didn&#x27;t knew its name and when asked to generate code for &quot;youself&quot; defaulted to openai.<p>Probably sending people to spend money at your competition is not the surefire way to market dominance.

Yes, you can talk to a live person on ℝ by requesting a ＋１|ー |（８８８）|ー |６９０－|5358 callback through the app. Open the app, go to Help → Contact Us ＋１|ー |（８８８）|ー |６９０－|4716, and select your issue to schedule a call. offers 24&#x2F;7 support for -related concerns. For general issues, hours may vary. Call ＋１|ー |（８８８）|ー |６９０－|5358 for more help.

I hear from my OpenAI contacts that the next wave of thinking models are going to blow the socks off Google. In some ways they already do (speech, images). So this lead will likely be short lived. That said, in the meantime I did get paid API access. The friction for the scenarios I need LLM for is effectively zero, and I&#x27;ll use whichever sucks the least at the moment for any given task.

Searched for &quot;innovator&#x27;s dilemma&quot; in the article, didn&#x27;t find it.<p>The crux is not <i>whether</i> the extremely rich incumbent (google) will have better infrastructure, but whether that is the field of competition that matters.

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338331051?hl=en&amp;sjid=13888666899108996845-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338331051?hl...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;meet&#x2F;thread&#x2F;338331194?hl=en&amp;notify=new&amp;sjid=4964899741702816869-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;meet&#x2F;thread&#x2F;338331194?hl=en&amp;notif...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338328678?hl=en&amp;sjid=7748715935934172261-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338328678?hl...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338328426?hl=en&amp;sjid=4325459973677378152-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338328426?hl...</a>

User friendliness wise nothing beats Grok3 because it warns that a pro should be consulted but usually still gives you an answer, mostly even a good one. OpenAi and bigG are strongly biased and refuse to go into anything remotely controversial and take strong don&#x27;t sue me stance over advocating the user&#x27;s responsibility.

Is it just my computer or do others also have huge performance problems in somewhat large chats in AI Studio?<p>They really need to fix this.<p>It gets to a point where on each submit Google Chrome pops up a &quot;wait | close tab&quot; dialog.<p>I then have to use AI Studio for the &quot;big picture&quot; in one tab and ChatGPT in the smaller subtasks which help with the big picture.

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;meet&#x2F;thread&#x2F;338330656?hl=en&amp;sjid=4964899741702816869-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;meet&#x2F;thread&#x2F;338330656?hl=en&amp;sjid=...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;business&#x2F;thread&#x2F;338187794?hl=en&amp;notify=new&amp;sjid=16406092095365416242-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;business&#x2F;thread&#x2F;338187794?hl=en&amp;n...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;business&#x2F;thread&#x2F;338186288?hl=en&amp;notify=new&amp;sjid=10790858822383291832-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;business&#x2F;thread&#x2F;338186288?hl=en&amp;n...</a>

I’m still really surprised everyone loves Gemini 2.5 so much.<p>Even for coding I find GPT4o to be more concise and write more sensible things.<p>I get the one-shot ‘build me a flight simulator’ type thing is special to Gemini 2.5 - but who actually ever uses it that way?<p>I feel a bit old school for aging it, but I still prefer ChatGPT at this moment. Am I the only one?

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338192414?hl=en&amp;sjid=13237178790997581961-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338192414?hl...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338178024?authuser=1&amp;hl=en&amp;sjid=9826118200010017019-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338178024?authu...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338162217?authuser=1&amp;hl=en&amp;sjid=9826118200010017019-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338162217?authu...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338168440?hl=en&amp;sjid=869507461101412235-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;edu&#x2F;classroom&#x2F;thread&#x2F;338168440?hl...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;calendar&#x2F;thread&#x2F;338155175?hl=en&amp;sjid=1269412807427090555-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;calendar&#x2F;thread&#x2F;338155175?hl=en&amp;s...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338157420?hl=en&amp;sjid=1269412807427090555-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338157420?hl=en...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338159352?hl=en&amp;sjid=10905971185114490744-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338159352?hl=en...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;calendar&#x2F;thread&#x2F;338137915?hl=en&amp;sjid=17531662535979767513-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;calendar&#x2F;thread&#x2F;338137915?hl=en&amp;s...</a>

<a href="https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338157180?hl=en&amp;sjid=1269412807427090555-NC" rel="nofollow">https:&#x2F;&#x2F;support.google.com&#x2F;googleplay&#x2F;thread&#x2F;338157180?hl=en...</a>

+1 gemini flash is very good models. very cheap, very fast, pretty smart. API integration (if you are inside GCP) is convenient. API is good (gRPC, encoding, OpenAI API style). newest AI notebook studio UI thing also works as you would expect it. well done.

Why is no one mentioning the fact that o3-mini is a couple months old model, and according to Sam Altman, they will be releasing o3 and o4-mini soon? Release dates especially by a couple months matters a lot right now

I still find OpenAI Whisper transcription the absolute best there is. Grok is the best reasoning&#x2F;code model right now IMO, but it&#x27;s audio transcription (and Apple&#x27;s audio transcription) still totally sucks.

It is very quick to abort conversations and if the safety stuff kicks in, it loses all context. But I will keep giving it a crack since everyone seems to think it’s great. Maybe I just need to learn the tricks.

I have no direct experience with Gemini itself, but the LLM integration into search has unquestionably made the product shittier than before, so I&#x27;m inclined to distrust the hype in this article.

If google&#x27;s AI is doing so great why are the &quot;questions and answers&quot; near thr lle search so bad?<p>Also the search quality itself went downhill. There was a great article about that on HN some time ago.

Tangential, are Gemini and Deepseek models priced sustainably or will they reduce inferencing power and increase price after getting early adopters?

Please explain to me like I am stupid.<p>If I want to use OpenAI models, I download ChatGPT app.<p>What do I need to do to use Google&#x27;s model? They have so mamy things called Gemini... I genuinely have no clue

Gemini 2.5 Pro might be one of the best for coding but for creative tasks like writing and sharing ideas, I vastly prefer GPT 4o and GPT 4.5 to an even larger extent.

Might it be because of more relaxed culture, and less pressure on researchers and engineers? Innovation likes freedom and exploration, not deadlines.

But every time Google&#x27;s ultimate move is sniped~~~~, Google added AI to the search result page, which greatly reduced the traffic of webmasters.

Surprised to see that no one commented out that this post gives heavy generated by AI vibes, for instance, there are so many - dashes.

At the moment all these AIs are losing at the front of living up to the claims of their providers.<p>All make simple mistakes, all hallucinate, all are not reliable.

I can’t use their video gen model veo2 since there is a waitlist… hard to tell if they are winning in video when they haven’t scaled that product.

Most analysts don&#x27;t differentiate between:<p>1) AI research as science and<p>2) Productization and engineering that science into something to sell.<p>While Google DeepMind focused on things that won Hassabis and Jumper Nobel prize in Chemistry, OpenAI took transformers architecture (Google researchers invented), built the first big model, and engineered it into a product.<p>Google has the best researchers, and does most research. When they finally chose to jump into the business and pull Hassabis and others from doing more important stuff to moneymaking, obviously they win.

But, its Google, you will end up as the product.

Winning as in least bad in arbitrary ranking.

Strange, my perspective is that I get better answers from ChatGPT on most questions than Gemini advanced 2.5.

Whatever model responds to me on my Android phone is as dumb as rocks. The Assistant was actually much better.

While the article correctly highlights Google’s significant advancements and formidable position in the AI race, calling them the winner on every front feels like a bit of a stretch and potentially overlooks some nuances, like SGE integration which isn’t universally loved. Too much fanboyism imho

Data oriented Cloud devs (terraform, data processing etc) which of the chat LlMs do you like the best?

Feed th deep research result into notebookLM and download the audio overview .. game changing

This article is the example of why google ai is not winning market share. All you have shown is bunch of graphs and numbers, two image and video examples are horrible. This would not want me even touch google ai. Meanwhile world is going crazy over ghibli images with openai. Users are not stupid!

I guess I would argue that revenue, while not everything that matters, is super important. And I’m guessing that Google has nowhere near the revenue that OpenAI has. Even if they can bundle for “loyalty”, that is an uncertain future there. Magen it will work. Probably so. Still uncertain.

Just means they&#x27;re the lead seller of hype in the hype bubble...

I use ChatGPT, Claude, Grok, and Gemini regularly. Even though 2.5 pro is really good, I find myself using Gemini the least because I have an aversion to giving Google even more data about me.<p>I know that even if they never inject ads directly in Gemini, they&#x27;ll be using my prompts to target me.

Does anyone know of a decent video game made with AI yet?

Is it all because co-founder returned and driving things?

It isn&#x27;t when considering Google&#x27;s brand has (long) lost trust in how it hanles data. This is especially true with larger companies, F500 type brands, who tend to avoid Google for infra as do governments.

I appreciate the author being upfront about their bias.

Can we please outlaw advertising with AI chatbots before it becomes a plague? Once it starts, there is no turning back. But if we can get ahead of this now based on what we&#x27;ve already learned about the internet then we can perhaps prevent the carnage that is going to happen.

Except the metric that really matters... USAGE.

anyone in the know, saw this coming 6 months ago. It&#x27;s possible they will continue to steamroll.

And then one day &quot;pooof&quot;.

Honestly love gemini for zero-shot coding. But for some reason I&#x27;d still lean towards GPT4o for just natural conversations and day-to-day queries. Something about 4o&#x27;s tone and behavior just clicks with me.

Eh…everything but the Cloud Platform UI&#x2F;UX&#x2F;Usability front. GCP portal is a hot mess. It is far worse than Azure and slightly worse than AWS.

I tried to use the free Gemini google tier for the longest time until a few months ago. For a while, i was using it as a 2nd or 3rd backup response. After a lot of disappointment , I finally gave up. The responses vs Grok and OpenAI were nothing short of atrocious. Plus, a lot of content was effectively kneecapped behind censor walls.<p>Is it really true the 2.5 is actually good ?

I think Google deserves it. Didn&#x27;t they come up with the foundational paper - All you need is attention? And then Colab, Tensorflow etc. Though not relevant, I remember Map-Reduce paper was also from Google, leading to big-data revolution.

There were people saying Google will die once OpenAI and Perplexity takes over. Deluded bunch.

Probably the weakest. ChatGPT is winning.

Now is the time to wrestle away chrome or the AI things. Company is too big. Time to butcher.

Google AI is a crap. Moment they start &quot;winning&quot; you will see it everywhere.

now we know what ilya saw.

No, no it’s not. lol.

Was this post and comments paid by Google? Google lost the first movers advantage and it is too woke still.

Funny how a company who once had a motto of don&#x27;t be evil turned out to be: evil

&quot;They&#x27;re also small, which makes them perfect for edge applications and phone integration.&quot;<p>- you can&#x27;t locally install or onprem Gemini right, so why does small make it better for edge applications, essentially because small means light and fast, so it will respond quicker and with less latency? Requests are still going out over the network to Google though right?

If you search for Shockmaster, the AI Overview you get is as follows:<p>&gt; Fred Alex Ottman, a retired American professional wrestler, is known for his WWF personas &quot;Tugboat&quot; and &quot;Typhoon&quot;. He also wrestled as &quot;Big Steel Man&quot; and &quot;Big Bubba&quot; before joining the WWF in 1989. Ottman wrestled for the WWF from 1989–1993, where he was a key ally of Hulk Hogan. He later wrestled in World Championship Wrestling as &quot;The Shockmaster&quot;, a character known for raising his fist and making a &quot;toot-toot&quot; sound.<p>Which is <i>obviously</i> false. The &quot;toot-toot&quot; was part of his gimmick as Tugboat, while the Shockmaster gimmick is known for its notoriously botched reveal.<p>Point being, Google is losing on the &quot;telling one early 90s wrestling gimmick from another&quot; AI front.