Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. The Return of the Utah Teapot – Siggraph 2026 [video](youtube.com ↗)
    discuss
  2. AI eyes in the sky: New satellites and AI transforming wildfire detection(theguardian.com ↗)
    discuss
  3. Sales calculation tool for Excel price lists
    discuss
  4. Qbix Server – PHP 100× faster than Nginx and PHP-fpm(github.com/qbix ↗)
    discuss
  5. The Input Layer(mg-crea.com ↗)
    discuss
  6. Show HN: Explore 2D semantic space with the Jev model(semanticspace.dev ↗)
    discuss
  7. King Charles has hesitations about AI(techcrunch.com ↗)
    discuss
  8. The Harms of Modern Lighting and the Fight to Bring Back Incandescent Bulbs(midwesterndoctor.com ↗)
    discuss
  9. Show HN: Using a diffusion model to write docs quickly(cortee.ai ↗)
    discuss
  10. Making a game for the GBA and PC from the same codebase(mattgreer.dev ↗)
    discuss
  11. lonely 20 questions, a game for people with no friends(foxmoss.com ↗)
    discuss
  12. Don't Trust the Super-App: A Case Study of Russia's Max(arxiv.org ↗)
    discuss
  13. Does this mean it's fine to use Claude subscriptions with third party harnesses?
    1comments
  14. New cat species identified for the first time in a century(cbc.ca ↗)
    1comments
  15. Moving Share Homes for Ukraine data's code to open source(mhclgdigital.blog.gov.uk ↗)
    discuss
  16. Grep your Claude Code/Codex history and jump back into the exact session
    discuss
  17. Hjfjhg
    discuss
  18. Building an Integrated Vector Database System in PostgreSQL(arxiv.org ↗)
    discuss
  19. Albanins discovering the long-banned car (1992)(washingtonpost.com ↗)
    1comments
  20. Backgrounding My Way to Japan – How to learn truth from fiction(aethermug.com ↗)
    discuss
  21. Show HN: Jawz – macro market state for AI agents, now a ChatGPT app (no login)(jawz.ai ↗)
    discuss
  22. Query Plan Rewriting in PostgreSQL(theconsensus.dev ↗)
    discuss
  23. If materialism is true, the United States is probably conscious(jstor.org ↗)
    discuss
  24. Deploying Monorepos Sucks(dominikkoch.dev ↗)
    discuss
  25. NeverLose – Crash-proof any Python function with one decorator(github.com/neverlose-io ↗)
    discuss
  26. The 2.5-hour AI-generated Odyssey movie is 2.5 hours too long(theverge.com ↗)
    discuss
  27. RepoGuard – CLI to stop Cursor and Copilot from bypassing architecture(github.com/taylormatematica-beep ↗)
    discuss
  28. Show HN: OpsGraph – evidence-first PostgreSQL investigations(opsgraph-site-seven.vercel.app ↗)
    discuss
  29. Show HN: I built a small tool to turn my essays into visual stories for YouTube [video](youtube.com ↗)
    discuss
  30. Are AIs Still Struggling with CAPTCHAs?(schneier.com ↗)
    discuss

Google illegally retains customer data,and I am taking legal action against them

48 pointsby 2h agomedium.com
12 comments
1h agoHN ↗

"But Google got caught precisely because of their own file‑restore robot, and this is my technical checkmate against them:"

Claudish (but not all of the text seems to be). Idk, I read this all day and am getting allergic to it. It's always the load-bearing decision worth my thought, and then some contradiction. It's not just this, it's THIS and THIS.

I've been using some GLM5.2 via Mistral lately and it doesn't have this... tone... But it makes some mistakes sometimes that Claude wouldn't.

Idk, just the idea that this was artificial expanded, makes me feel anyone could have written it, I could ask Claude for it directly. There will be no insight, even though there may be. What is happening to me, I am not anti AI.

54m agoHN ↗

I read this all day and am getting allergic to it. It's always the load-bearing decision worth my thought, and then some contradiction. It's not just this, it's THIS and THIS.

I kinda the mockery it makes of the "look how smart I am" internet comment types since it's basically their sort of linguistic drivel but cranked to 11.

1h agoHN ↗

Google’s own developer (Lalit_Kumar, 2025) wrote that the chat is actually not the JSON file in Drive, but only metadata. Due to the loss of the metadata, AI Studio “does not recognize” the chat — but this is not a deletion.

This is like deleting a shortcut and saying the file is gone, right?

1h agoHN ↗

That's not how I read the linked thread. The linked thread sounds like the JSON file also has metadata attached to it.

Since it's a JSON file, I wonder what's actually inside, and why the blog post doesn't mention opening it with a text editor.

1h agoHN ↗

If I go to court, I win. There is no other outcome.

Har har har. Spoken like somebody who's never been in a courtroom or read about a court case before. You still have to tick all the technicality boxes before you can win, starting with proving that the law actually says Google has to delete this. You also have to follow all the implicit rules of court procedure, which are the main reason you always need a lawyer.

58m agoHN ↗

Op doesn't realize there are so many legal terms that will make what he thinks is criminal not criminal and give them so many ways out of it these same people literally write the law good luck sir

44m agoHN ↗

To me the blog post feels as though AI has had significant input in the writing. Whether it's entirely AI written or AI ideated with some human input on the writing I don't know. So it's not really a surprise to see this sort of statement.

33m agoHN ↗

Lol at 'winning' a GDPR case against Google on a technical point. My first thought was 'Oh Sweet Summer Child' and realising that Max Schrems and NOYB aren't as well known as I originally thought.

https://noyb.eu/en https://schre.ms/

39m agoHN ↗

It’s odd that someone would assert that a legal action would be successful while citing no legal precedent.

All of the facts presented could be true, but what law is broken? What harm creates the basis for a civil suit?

29m agoHN ↗

This is great. I strongly encourage you go for it. Plenty of people have fought large corporations and received settlements. Get yourself a decent lawyer and consider starting up a class action!

Good luck. Despite what some people in this thread are saying, Google is definitely in the wrong here. What they will "lose" because of that, if anything, remains to be seen.