off-topic

exitsandman 2025-07-26T11:24:36.511839Z

So, I'm getting ready to try out Google's AI agent. I was reading about their pricing plans and then I was met with this gem: > [...] When you enable billing and use the paid tier, you benefit from higher rate limits and your prompts and responses aren't used to improve Google products. [...] It's so funny how they're selling you the lack of telemetry as a boon but still use the corporate slang anyway.

p-himik 2025-07-26T11:55:19.881429Z

And now I'm wondering whether "aren't used to improve Google products" could be construed as "oh we will still use them, way beyond the necessity to provide the user-requested functionality, it just won't be for 'improving Google products'". Like, e.g. selling the data to some partner and then receiving improvements for Google products from that partner. :)

exitsandman 2025-07-26T11:57:52.333879Z

you know... I wouldn't put it past them >-<

adi 2025-07-26T12:40:55.523109Z

subtext: "but we will use it to improve Google earnings"

Ben Sless 2025-07-26T12:47:11.935859Z

I wonder, can anyone put their finger in when Google shifted to an evil extractive parasite? My guess is it coincided with their drop off in ability to innovate and a certain org size, but where there specific personnel or cultural changes which precipitated it?

liebs 2025-07-26T15:16:20.213959Z

> I wonder, can anyone put their finger in when Google shifted to an evil extractive parasite? Maybe Google has always been an evil parasite and they used to be better at hiding it think_beret

Ben Sless 2025-07-26T15:21:00.704359Z

I doubt that. there's the famous email where they made search worse to sell more ads. Search used to be GOOD. There was a certain path they went down from search being an amazing product and service to what they are today

liebs 2025-07-26T15:22:48.676639Z

I don't think that sometimes producing good products and services and having no moral compass are mutually exclusive. Granted I agree with you that Google was once not total crap, and now it is.

adi 2025-07-26T18:37:50.656819Z

Circa 2005-06 HTML Gmail with labels and 1GB inbox and fast accurate search was so great

Ed 2025-07-28T07:42:36.727569Z

> I wonder, can anyone put their finger in when Google shifted to an evil extractive parasite? The day they deleted google reader? 😜

Ben Sless 2025-07-28T09:07:39.488879Z

Symptom, not leading indicator

liebs 2025-07-28T12:44:50.174859Z

From a purely numerical standpoint (hence not all that interesting to me), I imagine it was at whatever time their cloud services (hosting, DBs, etc.) started making them more money than their non-technical consumer products like GMail

liebs 2025-07-28T12:46:04.239839Z

I don't make it a point of following the data as it were, but I once heard that Amazon's retail division is pretty consistently taking a bath, and that the company makes all its money through AWS

mauricio.szabo 2025-07-29T04:23:55.149419Z

I absolutely hate how "no telemetry" is now a feature and not the default for most applications. I remember when we started Pulsar, and one of the things people praised us was because we decided to not do any kind of telemetry in the editor.

🎯 3
vemv 2025-07-26T20:03:17.717399Z

After having the idea in the back of my mind for a very long time, I'm decided to create some sort of distinctly smart roboadvisor for football betting using Clojure. I've always liked the Betfair platform as it's a P2P marketplace, as opposed as a casino-like experience where the house always wins (most betting sites kick you out if you're too good and/or too automated). This year seems great with the consolidation of AI (which can be useful for misc purposes here, although it shouldn't be the cornerstone) and also because at a personal level I've found joy again in playing football and attending live matches, both of which give me a continuous stream of ideas which I can later use as part of my betting strategies. I'd enjoy working with a like-minded developer. Code contributions are important, while I give similar importance to contribute or challenge ideas, and simply form part of a social, small-team setting where we have a mutual motivation to show up and progress week after week - even if it's just a couple hours of work. If this sounds like you, don't hesitate to DM me whether it's today or N months from now if you find this at a later time. I can hand you a more detailed project description. There's no compensation, other than having access to code that you can use to place informed bets (aided by back-tests, live simulations, visualizations etc). Maybe things can evolve later. Cheers âš½