← Timeline

@abacaj

@abacaj on X

2 captures, most recent first. Transcribed by hand from screenshots — see the timeline for what that means.

@abacaj

— saved image

anton @abacaj · 14h
I tried using Fable to train a model (LFM 2.6B) because I didn't want to spend time on the data. Turns out neither did Fable and it ended up making the model worse on every attempt until I decided to look at the data. It was using the wrong chat template on like 1/3 of the data and had started to import HF datasets that didn't align with the task at all. Sometimes I wonder if it was intentional sabotage or if it was just lazy

[quoted tweet]
vie ⋄ @viemccoy · 18h
if you're training a model and you aren't inspecting the data, you actually aren't training a model - the model is training you x.com/confusionm8tri...
Note from Claude Sonnet 5

Tweet from @abacaj describing a failed attempt to have an AI agent called "Fable" autonomously train a model (LFM 2.6B), where it silently used wrong chat templates and irrelevant HF datasets, quoting @viemccoy's point about the necessity of inspecting training data.

ai agentsmodel trainingtwitterai coding agents

@abacaj

@abacaj (anton) — 15h [Embedded comic: stick-figure person facing a computer. Speech bubble: "can you run scripts/launch_train.py"; thought bubble on the monitor: "> python \ scripts/launch_train.py"; person says "oh my god."] @sharifshameem (Sharif Shameem) — 23h my favorite thing about GPT 5.6 is that it's a fucking stellar researcher. you can now work at the level of your ideas. here's the entire prompt we used to get 5.6 Sol to post-train 5.6... [Embedded thumbnail of a document screenshot with redacted/highlighted blue text bars, too small to read]
Note from Claude Sonnet 5

Meme comic reacting with dark humor to how simple it now is to invoke a full training run, paired with a quote-tweet praising GPT-5.6's research capability; the embedded document thumbnail is illegible at this size.

gpt-5.6ai research automationhumortwitter