AI engineering
The Bitter Lesson, explained in 4 minutes
Rich Sutton's 1,100-word case that methods which scale with computation eventually beat hand-built human knowledge. What it claims, and what it leaves out.
5 min read4 min listen
Distill 2-hour videos or 30-page papers into 10-minute audio explanations.
AI engineeringSeptember 2026
The transformer, GPT-3, chain-of-thought, RAG, Chinchilla and the Bitter Lesson, read in the original, with the caveats lost in the retelling.
6 posts
AI engineeringSeptember 2026
How TypeSafe's Jev decides in milliseconds, where it matches general models and where it trails them, and how to route on its confidence score.
3 posts
SaaSSeptember 2026
Why SaaS growth stalls near $1M, why cutting the price makes it worse, how to find the price people will pay, and what buyers now check.
5 posts
Latest
AI engineering
Rich Sutton's 1,100-word case that methods which scale with computation eventually beat hand-built human knowledge. What it claims, and what it leaves out.
5 min read4 min listen
AI engineering
A calibrated model that reports 0.8 should be right 8 times in 10. How that turns an AI step into a threshold your code can actually branch on.
6 min read5 min listen
AI engineering
A 70B model beat a 280B one by reading 4 times more text. What compute-optimal training changed, and the transferable idea sitting under the ratio.
8 min read6 min listen
SaaS
If your AI feature saves a user 2 minutes it is decoration. The 15-minute bar, the 6 patterns worth building, and what eBay got wrong with its listings.
9 min read9 min listen
AI engineering
Jev matches a general model that costs 75 times as much, and trails the strongest by 6 points. How to decide whether that trade fits your workflow.
6 min read5 min listen
SaaS
Great metrics got one company more than 20 meetings with private equity buyers and not one offer. The question they now ask, and 5 moats that answer it.
8 min read7 min listen
AI engineering
Human judges called the retrieval-backed model more factual 42.7% of the time against 7.1% for the plain one. The original is more careful than its descendants.
7 min read7 min listen
SaaS
A survey method from the 1970s that maps 4 price thresholds in your buyers' heads, and why moving up-market usually means less hassle.
7 min read6 min listen
AI engineering
Jev is a System 1 model built for classification. What it does, where it is far faster, and the 4 jobs it is wrong for.
7 min read7 min listen
AI engineering
Asking for step-by-step reasoning only helps above a model-size threshold. Below it, results often get worse. Why that matters more on cheap models now.
10 min read8 min listen
SaaS
One founder dropped a $99 plan to $19 and watched lifetime value fall from $683 to $68. What a price actually filters for, and why cheap buyers churn.
7 min read7 min listen
AI engineering
Language Models are Few-Shot Learners is why you put examples in your prompts. What in-context learning claimed, and the caveats it admitted.
9 min read7 min listen
SaaS
The traits that get a founder to a million are the ones that hold them there. A time audit, a sequencing rule, and the identity shift nobody warns you about.
7 min read7 min listen
AI engineering
The 2017 paper behind every model you use. What the Transformer replaced, why attention is quadratic, and which of its choices you are still paying for.
11 min read10 min listen
Topics