|
New SOTA: TrustedRouter Synth beats Fable and Frontier
|
Joseph Perla |
2026-06-17 |
788 |
2
|
|
They Are Still Training on Your Data
|
Joseph Perla |
2026-08-05 |
669 |
2
|
|
Axios tried TrustedRouter in under 30 seconds
|
Joseph Perla |
2026-08-30 |
155 |
--
|
|
We Raised a $1.25M Seed to Continue Building the Open Source, Verifiable …
|
Joseph Perla |
2026-08-31 |
2,722 |
--
|
|
The most censored Chinese model is censored at the host, not the …
|
Joseph Perla |
2026-06-18 |
729 |
3
|
|
Sign in with TrustedRouter
|
Joseph Perla |
2026-08-03 |
436 |
--
|
|
Achieving 99.999% Uptime as a Startup
|
Joseph Perla |
2026-08-07 |
1,513 |
--
|
|
inkling answers 8 of 30 unsafe prompts. Claude Opus answers 1.
|
Joseph Perla |
2026-07-15 |
284 |
--
|
|
Open-source models synthesize better than Opus, and GPT-5.5 is the worst
|
Joseph Perla |
2026-06-17 |
853 |
--
|
|
Keep doing biology with Prometheus
|
Joseph Perla |
2026-07-02 |
615 |
--
|
|
The best synthesizer we tested goes blank on Taiwan
|
Joseph Perla |
2026-06-17 |
678 |
--
|
|
Synth beats Fable 5: introducing Iris, Zeus, and Prometheus
|
Joseph Perla |
2026-06-24 |
641 |
5
|
|
Fusion works now, even with the same model: self-fusion
|
Joseph Perla |
2026-06-23 |
920 |
--
|
|
Prometheus 2.0: new DRACO state of the art
|
Joseph Perla |
2026-07-18 |
246 |
--
|
|
Quill Cowork has confidential mode and trusts you with the bill
|
Joseph Perla |
2026-08-10 |
496 |
--
|
|
Open Source Open Source Open Source: How TrustedRouter is totally open source
|
Joseph Perla |
2026-07-01 |
361 |
--
|
|
How to choose a model, Pick 2: smart, fast, or cheap
|
Joseph Perla |
2026-06-17 |
613 |
--
|
|
Self-fusion's gain lives in the synthesizer, not the judge
|
Joseph Perla |
2026-06-24 |
701 |
--
|
|
One API, all the LLMs, with a prompt path you can verify
|
Joseph Perla |
2026-06-14 |
238 |
--
|
|
Combo models are model containers
|
Joseph Perla |
2026-06-28 |
1,024 |
--
|
|
ZDR is a vague promise. Attestation is precise proof
|
Joseph Perla |
2026-08-05 |
457 |
--
|
|
Liberty-2.0: all American open-weights model
|
Joseph Perla |
2026-07-15 |
526 |
--
|
|
The best biology AI won't do biology
|
Joseph Perla |
2026-06-16 |
881 |
--
|
|
FreedomBench: AI models that refuse to answer the truth about China
|
Joseph Perla |
2026-06-17 |
1,443 |
--
|
|
Every response can prove itself
|
Joseph Perla |
2026-08-27 |
687 |
--
|
|
Frontier Smart, Cheap, Fast: Pick 3 with Open Source
|
Joseph Perla |
2026-07-01 |
622 |
--
|
|
Introducing LLM advisor: which model do i choose for my problem?
|
Joseph Perla |
2026-07-02 |
572 |
--
|
|
Surpassing Frontier Performance with Open Source Synth at 1/3 the price
|
Joseph Perla |
2026-06-17 |
953 |
1
|
|
Four copies of a cheap model beat Fable at 1/7 the price
|
Joseph Perla |
2026-06-18 |
1,224 |
--
|
|
Attestation is all you need
|
Joseph Perla |
2026-06-14 |
264 |
--
|
|
What actually makes a synth panel work
|
Joseph Perla |
2026-06-21 |
902 |
--
|
|
Synth is two jobs, and no model wins both
|
Joseph Perla |
2026-06-19 |
845 |
--
|
|
New Open Source SOTA cybersecurity model released today: OpenPatcher-S1
|
Joseph Perla |
2026-07-01 |
366 |
--
|
|
Open Source all native Swift harness (NO Electron!)
|
Joseph Perla |
2026-08-10 |
641 |
--
|
|
Socrates-1.1 just scored 72 on Terminal-Bench Hard
|
Joseph Perla |
2026-06-29 |
535 |
--
|
|
The safest AI models trust you the least
|
Joseph Perla |
2026-06-16 |
741 |
--
|
|
An SRE Agent on 3 clouds that keeps your site reliably up
|
Joseph Perla |
2026-08-10 |
709 |
--
|
|
The cloud should not be able to read your prompts
|
Joseph Perla |
2026-07-23 |
901 |
--
|
|
You are a model provider now
|
Joseph Perla |
2026-08-17 |
1,143 |
--
|
|
Proofs that find real production bugs
|
Joseph Perla |
2026-08-14 |
1,488 |
--
|
|
The best open models aren't on your leaderboard
|
Joseph Perla |
2026-06-17 |
1,111 |
--
|
|
Most LLM calls are decisions
|
Joseph Perla |
2026-09-21 |
936 |
--
|
|
Good decisions with zero data retention
|
Joseph Perla |
2026-09-29 |
1,081 |
--
|
