It's time for developers and enterprises to build with Gemini Pro | Hacker News Reader

It's time for developers and enterprises to build with Gemini Pro

developers.googleblog.com

18 pointspiecerough19 comments

I've been burned too many times by Google's need to constantly kill its own products. I strongly caution anyone from tying their business to a Google product until the product has existed for more than a decade. Google's strategic ADHD does not instill confidence when combined with the knowledge that these AI-models are very expensive to run.

Very true. I bought an Epson Ecotank printer a while ago that had on the box, "Google Cloud Print" and I used it and it was amazing, I could print a photo I just took, to my home printer, from anywhere in the world. Then they killed it, and my useful printer stopped being so useful.

I don't trust Google anymore.

I don't understand this perspective, yet I keep seeing it in these Gemini threads. The Google Graveyard is large, yes, but they've considered themselves an "AI-first company" for several years now. It's clear by now that consumers will expect it to be part of search as well as part of their smartphone experience. That's Google's core business, not some random app or moonshot. Do you really think Google's going to say in a few years, "well, this AI stuff was neat but we're going to sunset it now"?

The question is will they say "well, this model was great but we have better/shinier models now, so we'll sunset it. Please change all your software to use the new thing with a new API interface and different expected results".

This exactly. If you're ever considering using a google product for something important, you would do well to read this classic blog post first.

https://steve-yegge.medium.com/dear-google-cloud-your-deprec...

---

Dear RECIPIENT,

Fuck yooooouuuuuuuu. Fuck you, fuck you, Fuck You. Drop whatever you are doing because it’s not important. What is important is OUR time. It’s costing us time and money to support our shit, and we’re tired of it, so we’re not going to support it anymore. So drop your fucking plans and go start digging through our shitty documentation, begging for scraps on forums, and oh by the way, our new shit is COMPLETELY different from the old shit, because well, we fucked that design up pretty bad, heh, but hey, that’s YOUR problem, not our problem.

We remain committed as always to ensuring everything you write will be unusable within 1 year.

Please go fuck yourself,

Google Cloud Platform

I did have some nagging sense that what i was writing was very unoriginal :)

You can always find it again by searching <google cloud fuck you> in your favorite search engine ;)

> Do you really think Google's going to say in a few years, "well, this AI stuff was neat but we're going to sunset it now"?

AI, no; this AI stuff, almost certainly. It's kind of like arguing that Google's never going to give up on chat - sure, chat is important, but that's not exactly helpful to the devs who made use of the XMPP interop in Google Talk to build great apps that then broke when Google killed it.

I would love to develop on this, and onboarding looks pretty straightforward and pain free.

However with Googles history of junking projects, I just wouldn’t have the confidence to take this any further than a hobbyist tinkering.

It is difficult to use Gemini because they don't seem to have a public equivalent to chat.openai.com yet. But there is https://bard.google.com but I can't verify the model used.

Either way, I uploaded this picture from the Gemini example prompt library to both Bard and GPT4 and asked "How much does each item cost?". Bard showed me a map of Bali and GPT4 got every item correct.

https://imgur.com/a/mymY23g (my test)

Not sure if Bard is using Gemini yet, but them going around hyping up Gemini without a working public prototype seems like a fail on the statement "you only get one chance to make a first impression".

ChatGPT spread by word of mouth from people using it. Google is taking the opposite approach it seems.

> I uploaded this picture from the Gemini example prompt library to both Bard and GPT4 and asked "How much does each item cost?". Bard showed me a map of Bali

That’s actually hilarious and would fit perfectly in a Silicon Valley episode.

Isn't the deal with this post about the "Google AI Studio" they released today? It's different than Bard. Also, I don't see the point in comparing Gemini Pro to GPT-4, since Google's own paper describes it as only slightly better than GPT 3.5.

Unless it can generate billions of ad revenue for Google, or is used internally by Google to generate billions in ad revenue, it will be scrapped or otherwise ruined in service of generating billions in advertising revenue, so, why?

It is quite impressive, and it will not shy away from having to do simple arithmetic, unlike GPT4! But man the hoops you have to go through to use it if you are not in the list of approved countries...

Meanwhile GPT4 just writes code for arithmetic and runs it which is far more reliable

It's time for Google to start behaving like a trustworthy partner both in terms of what their marketing team says but also management commitment to long term support of their products without rug pulls every other year.

Then "developers and enterprises" will come. ( more the enterprises because devs have a slightly better strategic mind than a goldfish )

Where's the audio capability of Gemini Pro? The only feature that made their model unique and they removed it. Embarrassing.

Look this is all nice but when you go back to the incentives there’s really good reasons to not build with your stuff Google.

You see, we know every time we use AI instead of one of your existing products, you lose money. A lot of money in case of anything that’s monetized like search. So you’re in a bind - you need people to believe you own the AI future but not actually use it.

AI for you is the bright flare you want investors to look at so they don’t fully understand how much the ads business is burning under your layers of smoke screens and unsustainable metrics juicing (How many more times do you think you can raise Youtube fees and close products like Google Domains?).

Your own company dysfunctional. Fear and terror about layoffs cause people to optimize for kpis that shift every month. Nobody wants to do the right thing because that’ll get you fired.

You feel compelled to cherry pick metrics and compare non released models along those cherry picked metrics to existing and working competition products. You make fake videos to get headlines to distract people from how far you are behind because every agressive move you can make in AI will cannibalize your existing business. Your own pricing communication compares Gemini to GPT 4 to look favorable when it’s available performance is more GPT 3.5 territory. Why?

A capable leader would have done the hard thing, but you have no leaders only managers who managed to survive the VP backstabbing gauntlet by diplomacy. They are unable to enact the decisive change needed, unable to sell to the shareholders that you can’t have both things - booming ads product and AI business because Microsoft shrewdly shanked you with their OpenAI shiv and forced a consumer battle you were not ready for.

You sit on proprietary technology, getting hammered by OpenAI on the top and eaten alive by LLamas on the bottom. There’s no space for a closed source #3 in this market, let alone one who had grown fat and untrustworthy in this market.

We all can see you are bleeding - your once proud search result page which didn’t need people to even scroll, let alone go to page 2 now has been revamped to infinite scroll as you are in full value extraction mode. Your incentives are now to kill any consumer value that may still be found in your products to survive the next quarter, and the next. And because you are failing, you still have 10% layoffs every quarter too, which cripples your organization.

You are not well google; you are selling your souls; firing your employees and squeezing your customers to make your shareholders right quarter after quarter. We know you have all the elements needed to run a great AI business, but unfortunately for you, most of a great AI business is not incremental but competitive with your own ads financed products. You failed to align your own organization on AI Transformation - instead you left your teams rudderless and without guidance and it shows.

In short, you are a wounded giant leaping from short term bandaid to short term bandaid, resorting to lies and smokescreens to appear in control. You’ve known peace and control for too long to react in wartime. Just like your investors, your “leaders” still believe they deserve that stock bonus today if they just juice to meet the metrics, what happens tomorrow be damned.

Even your lawyers and executives got so sloppy, they openly ignored the law and deleted evidence (wtf Google?), making Epic take a piece out of you where they couldn’t from Apple.

There’s blood in the water - your blood and everyone in this Red Ocean understands it’s yours. Meta, Microsoft, OpenAI and Apple, they all are plunging their knives into you. And you are flailing. Only a fool would build on your business offerings in AI right now, your crazed moves are too erratic, too high risk.

Go away google, you are drunk. Let’s talk again in a year when you are bled to size.

Looks like it's launching with some nice developer tools: low-code/no-code AI agent workflow, a prompt playground, etc.

Reply on news.ycombinator.com