frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Google AI Studio's "Delete" button is misleading

https://www.youtube.com/watch?v=LBQfNbPCxQw
1•Bitu79•52m ago

Comments

Bitu79•52m ago
I know you changed the button based on sentences you copied from my posts. And I have also proven that your backend keeps the data even after emptying the trash, which means you are illegally retaining user data.

Google AI Studio’s “Delete” Button Is Not Real Deletion — This Is Illegal Data Retention in the EU

I have proven that Google AI Studio does not delete user data, even when the interface claims it does. This is not a UI bug. This is undisclosed backend behavior and illegal data retention under EU law. When you press Delete, Google only removes the local JSON file from the client side. The server-side session, context, and conversation data remain fully intact. Even after emptying the trash, the backend still keeps everything. When you restore the JSON, the chat continues from the exact same internal state, which is only possible if the server never deleted anything. This is the equivalent of saying: I have a key and a house. The house disappears, but I can still walk through the door. If the data had truly been deleted, the model would not be able to continue the conversation. But it does, which proves the backend never removed the data. Under GDPR, the right to erasure, data minimization, purpose limitation, lawful basis, and transparency are all violated. Google claims the data is deleted, secretly retains it, and continues to process it. This is not a misunderstanding. This is not a UI issue. This is undisclosed internal behavior and unlawful data retention. I proved it with a simple logical test that anyone can reproduce.

Official Google emails about the recovery process:

https://archive.org/details/gmail-0-1805000041398-confirmati...

I prove the authenticity of the video by saving the long screenshot from the Google My Activity page, the PDF‑exported screenshot, and the full HTML archive into both my Google Drive and the Internet Archive. These files can be compared with the events shown in the video down to the exact minute. Google My Activity logs confirm every browser action, and the saved files are fully synchronized with the video in both timing and content.

Google My Activity:

https://drive.google.com/file/d/1Vywb6sim4lVin3d4dA9CUQMF_Pr...

https://archive.org/details/my-activity_202607

Part 2 of the video can be viewed here:

https://youtu.be/m4U6ajKuLRo

Google AI Studio’s “Delete” button is misleading.(Part 3):

https://youtu.be/8Pg-7eNPUWM

Using AI with a Critical Eye

https://www.loserbydesign.com/new-page-3
1•yehiaabdelm•1m ago•0 comments

Solving Fermat: Andrew Wiles

https://www.pbs.org/wgbh/nova/proof/wiles.html
1•1970-01-01•4m ago•0 comments

The Myth of the 'Waterfall' SDLC

http://www.bawiki.com/wiki/Waterfall.html
1•fanf2•4m ago•0 comments

Show HN: SilkDrop – MilkDrop port for web with live editor

https://silkdrop.vercel.app/
1•usernameis42•6m ago•0 comments

Experts warn current Starship heat shield tech is a "dead end" for rapid reuse

https://arstechnica.com/space/2026/07/despite-recent-successes-rapid-reuse-of-starship-remains-a-...
1•greedo•7m ago•1 comments

Softmax Approximation for FPGAs Using Taylor Series and Pade Approximants

https://leetarxiv.substack.com/p/approximating-softmax-for-fpgas
1•birdculture•7m ago•0 comments

Pivotick is network graph library to facilitate pivoting

https://pivotick.github.io/Pivotick/
1•adulau•7m ago•0 comments

Show HN: Latest Q2 AI SaaS Trends Report

https://relvehq.com/ai-reports/2026-q2-the-state-of-ai-for-saas-companies-fullview
1•hamza_ali_shah•9m ago•1 comments

README, not

https://blog.yossarian.net/2026/07/16/README-not
3•lr0•11m ago•1 comments

Sam Altman says we are in the singularity: 'This is the moment'

https://www.businessinsider.com/sam-altman-openai-the-singularity-agi-prediction-anthropic-nvidia...
2•doener•11m ago•1 comments

Nvidia in Talks with OpenAI to Guarantee $250B Financing for Data Center

https://www.wsj.com/tech/ai/nvidia-in-talks-with-openai-to-guarantee-250-billion-financing-for-da...
1•u1hcw9nx•11m ago•0 comments

Claude shared chats and Artifacts may have ended up on Google

https://techcrunch.com/2026/07/27/psa-your-claude-shared-chats-and-artifacts-may-have-ended-up-on...
2•speckx•18m ago•0 comments

Waymo crashes 1/3 as much as a human driver, says IIHS – with some caveats

https://electrek.co/2026/07/25/waymo-is-2-3-safer-than-a-human-driver-says-iihs-with-some-caveats/
2•Bender•19m ago•0 comments

Octane – compiled and faster React from creator of Inferno

https://octanejs.dev
1•swe_dima•19m ago•0 comments

E-bike bans in schools highlight the growing debate over freedom vs. safety

https://electrek.co/2026/07/27/e-bike-bans-in-schools-highlight-the-growing-debate-over-freedom-v...
1•Bender•22m ago•0 comments

I mapped how 140 governments legally treat Bitcoin

https://thebitcoinact.xyz/bitcoin-legal-map
1•The_Bitcoin_Act•23m ago•0 comments

Show HN: Nexmy – 100% offline Android finance tracker with local AI

https://nexmy.app/
1•BlackSharkDev•25m ago•0 comments

Show HN: A board game with opponents trained by self-play

https://mighty840.itch.io/samudra-manthan
1•sharang33•26m ago•0 comments

Show HN: Full complex text shaping and rendering on ESP32 with 320KB usable RAM

https://github.com/waruyama/flattype-cyd-demo
2•jansan•28m ago•1 comments

A political compass for AI where anyone can add their stance

https://theaicompass.io/?preview=default
1•uncommoncrawl•28m ago•0 comments

From ADHD chaos to IMPOSSIBLY productive and chill?

https://www.taskloco.com/
1•taskloco_nyc•30m ago•0 comments

RevelRaw – a Mac RAW editor that looks at the photo before suggesting presets

https://revelraw.com/
4•kernel_init•30m ago•0 comments

Dozens sickened in 17 states recalled eggs feared contaminated deadly bacteria

https://www.dailymail.com/health/article-16008335/dozens-sickened-recalled-eggs-salmonella.html
2•Bender•30m ago•0 comments

ThinkFull: A 22-specialist Claude Code skill, not an agent pipeline

https://github.com/SuperNotesOnline/ThinkFull
1•jazziejoy•30m ago•0 comments

Forth

https://xkcd.com/3277/
2•beardyw•31m ago•0 comments

Platform engineering 2.0 mitigates AI security and compliance risks

https://platformengineering.org/blog/how-platform-engineering-2-0-mitigates-ai-security-and-compl...
2•CrankyBear•32m ago•0 comments

UK's iconic Jodrell Bank telescope faces axe in science cuts

https://www.bbc.co.uk/news/articles/c5y6r9rdxygo
1•mellosouls•33m ago•1 comments

Anthropic changed a Claude Code default–how I insulated my framework

https://mauroepce.dev/blog/trust-boundaries-claude-code
1•oalders•33m ago•0 comments

Going Beyond Zero: A New Paradigm for Enterprise Security

https://blog.google/security/going-beyond-zero-a-new-paradigm-for-enterprise-security/
2•denysvitali•34m ago•0 comments

Fwupd 2.1.7 Released with Many Bug Fixes

https://www.phoronix.com/news/Fwupd-2.1.7-Released
2•Bender•35m ago•0 comments