Tom Jones

Bootstrapped solo founder. Building Tīrtha, a cost efficient local AI engine with memory and verification for reliable long horizon agents.

Rock Hill Sc. Joined on Jun 27, 2026 https://tirtha.ai/

Work

Founder and CEO of Tirtha.ai

Tom Jones

Aug 9

Microsoft measured our thesis, and we still cannot quote ours

#ai #testing #benchmarking #programming

7 min read

Want to connect with Tom Jones?

Create an account to connect with Tom Jones. You can also sign in below to proceed if you already have an account.

Create Account

Already have an account? Sign in

Tom Jones

Aug 7

A stable aggregate is not a stable measurement

#benchmarking #testing #ai #datascience

4 min read

Tom Jones

Aug 7

We stopped choosing models and started choosing what to measure

#ai #llm #benchmarking #engineering

8 min read

Tom Jones

Aug 2

I built a page that only publishes numbers surviving our own checks. It caught me inside the hour.

#showdev #ai #testing #webdev

5 min read

Tom Jones

Jul 29

We built a router to predict when a cheap model is enough. It does not work.

#ai #machinelearning #llm #devops

4 min read

Tom Jones

Jul 26

Tokens Are Not the Unit

#ai #llm #programming #architecture

12 min read

Tom Jones

Jul 26

Your company already runs the hardest parts of an AI agent platform

#ai #llm #architecture #productivity

9 min read

Tom Jones

Jul 5

Everyone is hardening the structure. Nobody is passing down the why.

#ai #llm #machinelearning #agents

3 min read

Tom Jones

Jul 4

I didn't mean to build this. Looking for testers anyway.

#ai #webdev #opensource #startup

2 min read

Tom Jones

Jul 2

I got my coding agent to tie the frontier for about 8x less. Here is the honest benchmark.

#showdev #ai #machinelearning #llm

3 min read

Tom Jones

Jun 29

Serving cheap when two models agree: a measured cost lever

#testdev #ai #llm #infrastructure

5 min read

Tom Jones

Jun 28

Frontier-Quality Coding at Cheap-Tier Cost: What We Built, and How We Measured It

#ai #testing #llm #devops

5 min read

Tom Jones

Jun 28

The Two-Channel Problem: Structure and Soul for Reliable Long-Horizon Agents

#ai #llm #machinelearning #agents

4 min read

DEV Community

Tom Jones

Badges

4 Week Community Wellness Streak

2 Week Community Wellness Streak

1 Week Community Wellness Streak

Writing Debut

Microsoft measured our thesis, and we still cannot quote ours

Want to connect with Tom Jones?

A stable aggregate is not a stable measurement

We stopped choosing models and started choosing what to measure

I built a page that only publishes numbers surviving our own checks. It caught me inside the hour.

We built a router to predict when a cheap model is enough. It does not work.

Tokens Are Not the Unit

Your company already runs the hardest parts of an AI agent platform

Everyone is hardening the structure. Nobody is passing down the why.

I didn't mean to build this. Looking for testers anyway.

I got my coding agent to tie the frontier for about 8x less. Here is the honest benchmark.

Serving cheap when two models agree: a measured cost lever

Frontier-Quality Coding at Cheap-Tier Cost: What We Built, and How We Measured It

The Two-Channel Problem: Structure and Soul for Reliable Long-Horizon Agents