frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Open in hackernews

Ask HN: How to ask questions to LLMs privately?

2•quietthrow•15h ago
It’s an incredible time we live in. We can ask any question we want the previously we may have been embarrassed to ask and get an answer and learn and improve ourselves daily. Just one thing I can’t wrap my mind around is privacy. We all seem to be asking sometimes stupid and some times incredibly personal questions to these llms. Questions that we may not even speak out loud from embarrassment or shame or other such emotions to even our closest people. How are these companies using our data ? More importantly what are you all doing to protect yourself from misuse of your information? Or is it if you want to use it you have to give up such privacy and uncomfortableness ? What are some solutions available today that people are using there are somewhat privacy, preserving or fully privacy preserving?

Comments

wredcoll•15h ago
The same as its always been: run your own.
thoroughburro•15h ago
To give folks more of a toehold: LM Studio is a cross platform interface that makes it relatively easy to play with local models.

https://lmstudio.ai

mindcrime•14h ago
Self host using open models. There are multiple ways to do it but one pretty straightforward way is to run Ollama[1] and - if you want web chat - OpenWebUI[2].

[1]: https://github.com/ollama/ollama

[2]: https://github.com/open-webui/open-webui

trilogix•11h ago
Check HugstonOne. The definition of privacy.

The Hype is the Product

https://rys.io/en/180.html
1•lr0•23s ago•0 comments

Most developers use AI in their daily workflows – but they don't trust it

https://www.zdnet.com/article/most-developers-use-ai-daily-in-their-workflows-but-they-dont-trust-it-study-finds/
1•CrankyBear•34s ago•0 comments

Show HN: Claude Loop – defined plan execution over multiple iterations

https://github.com/DeprecatedLuke/claude-loop
1•kachapopopow•1m ago•0 comments

Is Roman concrete more sustainable? It's complicated

https://arstechnica.com/science/2025/07/is-roman-concrete-more-sustainable-its-complicated/
2•ulrischa•2m ago•0 comments

Show HN: JavaScript algorithm prep that teaches language-specific patterns

https://www.memoized.io/
2•elduderino05•3m ago•0 comments

Descent of Inanna into the Underworld

https://en.wikipedia.org/wiki/Descent_of_Inanna_into_the_Underworld
1•alganet•3m ago•0 comments

Ask HN: Why do so many developers dislike C when I find it inspiring?

1•silentpuck•4m ago•0 comments

SimpleX Chat v6.4.1: welcome your contacts, review members to protect groups

https://simplex.chat/blog/20250729-simplex-chat-v6-4-1-welcome-contacts-protect-groups-app-security.html
2•Sami_Lehtinen•4m ago•0 comments

Evidence for hepatitis C virus involvement in bipolar disorder and schizophrenia

https://www.nature.com/articles/s41398-025-03387-3
2•PaulHoule•8m ago•0 comments

Hask Is Not a Category

https://math.andrej.com/2016/08/06/hask-is-not-a-category/
1•fanf2•9m ago•0 comments

60k Gazans have been killed. 18,500 were children. These are their names

https://www.washingtonpost.com/world/interactive/2025/israel-gaza-war-children-death-toll/
12•rendang•11m ago•1 comments

Show HN: VoxChat – Anonymous Voice Chat Platform

https://voxchat.me/
1•huhtho•12m ago•0 comments

Northwest Mekeo Lacks Coronal Phonemes [pdf]

https://julietteblevins.ws.gc.cuny.edu/files/2016/10/Blevins2009e-UniversalBitesDustMekeo.pdf
1•thunderbong•12m ago•0 comments

BitVMX New Open Source Components: Key Management, Storage, and Configuration

https://bitvmx.org/knowledge/introducing-bitvmx-new-open-source-components-key-management-storage-and-configuration
1•wslh•14m ago•0 comments

Show HN: How to get instant advice from YC without needing to get in

https://www.alorapodcasts.com/chat/ycombinator
1•nkabbani10•14m ago•0 comments

Interacting with Text Adventures Through Perl

https://entropicthoughts.com/interacting-with-text-adventures-through-perl
2•Bogdanp•17m ago•0 comments

Nvidia-backed startup invents Ethernet memory pool – up to 18TB of DDR5

https://www.tomshardware.com/tech-industry/nvidia-backed-startup-invents-ethernet-memory-pool-to-help-power-ai-claims-it-can-add-up-to-18tb-of-ddr5-capacity-for-large-scale-inference-workloads-and-reduce-per-token-generation-costs-by-up-to-50-percent
1•pulse7•17m ago•0 comments

Paninaro

https://paninaro.onrender.com
1•immercato•19m ago•2 comments

How do I get ChatGPT to forget about me?

2•anupshinde•22m ago•0 comments

Show HN: I made "iChatroom" a dark webish type chatroom

https://www.ichatroomapp.com/
1•teamunited•23m ago•0 comments

Agent Interaction Guidelines

https://linear.app/developers/aig
1•GarethX•23m ago•0 comments

Flaw in Gemini CLI coding tool could allow hackers to run nasty commands

https://arstechnica.com/security/2025/07/flaw-in-gemini-cli-coding-tool-allowed-hackers-to-run-nasty-commands-on-user-devices/
1•llamasushi•23m ago•0 comments

End times of Intel [video]

https://www.youtube.com/watch?v=4ADwOm7Z3aI
1•vnorilo•25m ago•0 comments

The trade loophole that's costing the UK economy billions [video]

https://www.youtube.com/watch?v=jtqw3N4NZcA
1•ksec•27m ago•0 comments

Show HN: We built an internal MCP to improve Plaid engineering's productivity

https://plaid.com/blog/plaid-internal-mcp-server-productivity/
1•allenschen•27m ago•0 comments

One-and-done HIV protection in infants

https://www.nature.com/articles/s41586-025-09330-2
2•labanimalster•28m ago•1 comments

AMD Threadripper 9980X and 9970X Linux Benchmarks

https://www.phoronix.com/review/amd-threadripper-9970x-9980x-linux
4•Bezod•28m ago•1 comments

Fast

https://www.catherinejue.com/fast
4•gaplong•29m ago•0 comments

India overtakes China as biggest smartphone exporter to the United States

https://www.cnn.com/2025/07/29/business/india-china-us-smartphone-supplier-intl
6•LopRabbit•30m ago•0 comments

RunReveal Raises $7M Seed to Build the AI-Native Security Data Platform

https://blog.runreveal.com/runreveal-raises-7m-ai-security-platform/
1•ejcx•31m ago•0 comments