HNHacker News
TopNewBestAskShowJobs

ljlolel

2,702 karma · joined March 10, 2007

http://www.jperla.com/
submissionscomments
ljlolel··on Stripe will reportedly acquire OpenRouter for $7B+
we added that to our router https://trustedrouter.com/sign-in-with-trustedrouter
ljlolel··on Stripe will reportedly acquire OpenRouter for $7B+
what are you looking for in an alternative?

my site is the only end to end encrypted one TrustedRouter.com

ljlolel··on Stripe will reportedly acquire OpenRouter for $7B+
when you use a closed source router you don’t know what it’s doing.

that's why my site TrustedRouter is hosted and end to end encrypted and you can actually see what it’s doing.

ljlolel··on Stripe will reportedly acquire OpenRouter for $7B+
what are you looking for in a competitor?

doing research for my end to end encrypted router TrustedRouter

ljlolel··on DeepSeek API Pricing Update
it says they actively train on your data
ljlolel··on DeepSeek V4 Pro 0813
depends on your threat model
ljlolel··on DeepSeek V4 Pro 0813
goes all the way to keys owned by Nvidia and Intel
ljlolel··on Show HN: 100% native Swift harness (NOT Electron)
yes

should be able to byok or add it if not. it’s open source

ljlolel··on Codex in ChatGPT desktop app for Linux is now in preview
that’s why i made a clone of macos libraries to compile SwiftUI to linux to make something like this faster (for non Electron!)
ljlolel··on My Agent Setup
Keep the memory/tool silo in Hermes. I run TrustedRouter; trustedrouter/zdr gives that agent a ZDR-only model hop, but it cannot make Codex or Claude subscriptions ZDR. https://trustedrouter.com/blog/one-api-all-llms-provably-pri...
ljlolel··on DeepSeek V4 Pro 0813
No, not any
ljlolel··on DeepSeek V4 Pro 0813
With confidential compute / TEEs you can guarantee that the code is running, it's verifiable with remote attestation
ljlolel··on DeepSeek V4 Pro 0813
5% instead of 5.5%
ljlolel··on DeepSeek V4 Pro 0813
should use an open source one you can audit like my site TrustedRouter
ljlolel··on DeepSeek V4 Pro 0813
it's confidential compute, it's open source and you can verify yourself that it's not reading the prompts
ljlolel··on DeepSeek V4 Pro 0813
it's not open, I know people rejected by them (then they went to my site to be listed Trustedrouter.com)
ljlolel··on DeepSeek V4 Pro 0813
TrustedRouter is hosted and full opensource!
ljlolel··on DeepSeek V4 Pro 0813
I built TrustedRouter so this can fail closed. min_privacy=zdr rejects the request when the model has no ZDR provider; confidential requires provider-side confidential compute. https://trustedrouter.com/blog/how-confidential-computing-pr...
ljlolel··on Show HN: 100% native Swift harness (NOT Electron)
thanks!
ljlolel··on Show HN: 100% native Swift harness (NOT Electron)
i’m pretty strong with swift and i wanted other good engineers with good taste to make something
ljlolel··on DeepSeek V4 Flash 0731
I do but that's become my work is routing between all the models and end to end encryption, and making new models from these models https://trustedrouter.com/blog/synth-iris-prometheus-zeus
ljlolel··on Everyone is building LLM routers, we deprecated ours
nobody else is putting a router inside a Trusted Execution Environment https://trustedrouter.com/
ljlolel··on Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it
so interesting!!
ljlolel··on Using an open model feels surprisingly good
yes but AllRouter does automatic failover and can burst to cloud as well or use local when needed
ljlolel··on Kimi K3 Now Available via Telnyx Inference API
This (Telnyx) is available on TrustedRouter.com
ljlolel··on Kimi K3 Now Available via Telnyx Inference API
This (Telnyx) is available on TrustedRouter.com
ljlolel··on Using an open model feels surprisingly good
you can use claude code or codex with open models using AllRouter https://github.com/Lore-Hex/AllRouter
ljlolel··on Kimi-K3 Releases on HuggingFace 7/27
can send safely context if there’s confidential computing ala my site https://trustedrouter.com/
ljlolel··on Show HN: Echo – Fable-level results at 1/3 the cost using open-weight models
everything has been updated
ljlolel··on Show HN: Echo – Fable-level results at 1/3 the cost using open-weight models
ensemble models always did the best at Kaggle

we did the same: https://trustedrouter.com/blog/prometheus-2-new-draco-state-...

← PreviousPage 2 of 30Next →