Stories by kcorbitt

Codex, File My Taxes. Make No Mistakes

A Pocket Guide to Surviving the Robot Apocalypse

Show HN: RULER – Easily apply RL to any agent

Everything I know about reward hacking

Show HN: ART – a new open-source RL framework for training agents

ART·E: how we built an email research agent that beats o3

Using GRPO to Beat o1, o3-mini and R1 at “Temporal Clue”

Analyzing OpenAI's Reinforcement Fine-Tuning: Less Data, Better Results

Using reinforcement learning and $4.80 of GPU time to find the best HN post

Show HN: Agent.exe, a cross-platform app to let 3.5 Sonnet control your machine

DPO fine-tuning outperforms SFT

OpenPipe Mixture of Agents: Outperform GPT-4 at 1/25th the Cost

What we've learned in 3 days of Llama 3

Mixtral Curious? Comparing Mistral 7B and Mixtral for fine-tuning

S-LoRA: Serving Thousands of Models from One GPU for Fun and Profit

Is AI the next crypto? Insights from HN comments

Fine-tune your own Llama 2 to replace GPT-3.5/4

Show HN: Automatically convert your GPT-3.5 prompt to Llama 2

TaxyAI: Open-source browser automation with GPT-4

Tell HN: YC will help you find a co-founder

1