1
LLMs reward expertise
In the 2010s, if you had technical gaps (say, you couldn’t write CSS), you had to either rely on a skilled colleague or just hope that the answer to your exact…
Sean Goedecke
Sean Goedecke's personal blog
- Popular posts
- 46
- Total points
- 16,344
- HN comments
- 8,823
- Rank
- #40 personal
Visit seangoedecke.com → RSS feedOn Hacker News Oct 2020 – Nov 2026
Career & Management 22AI & Machine Learning 17Software Engineering 16Tech Industry & Business 9Society, Policy & Culture 7Life, Learning & Writing 5
2
What's the strongest AI model you can train on a laptop in five minutes?
What’s the strongest model I can train on my MacBook Pro in five minutes? I’ll give the answer upfront: the best 5-minute model I could train was a ~1.8M-param…
3
OpenAI's new open-source model is basically Phi-5
OpenAI just released its first ever open-source large language models, called gpt-oss-120b and gpt-oss-20b. You can talk to them here. Are they good models…
4
Why DeepSeek is cheap at scale but expensive to run locally
Why is DeepSeek-V3 supposedly fast and cheap to serve at scale, but too slow and expensive to run locally? Why are some AI models slow to respond but fast once…
5
DeepSeek-V4-Flash means LLM steering is interesting again
Ever since Golden Gate Claude I’ve been fascinated with “steering”: the idea that you can guide LLM outputs by directly manipulating the activations of the…
6
How I use LLMs as a staff engineer
What I use them for and what I don't
7
Is chain-of-thought AI reasoning a mirage?
Reading research papers and articles about chain-of-thought reasoning makes me frustrated. There are many interesting questions to ask about chain-of-thought…
8
Two different tricks for fast LLM inference
Anthropic and OpenAI both recently announced “fast mode”: a way to interact with their best coding model at significantly higher speeds. These two versions of…
9
If you are good at code review, you will be good at using AI agents
Using AI agents correctly is a process of reviewing code. If you’re good at reviewing code, you’ll be good at using tools like Claude Code, Codex, or the…
10
Diffusion models explained simply
Transformer-based large language models are relatively easy to understand. You break language down into a finite set of “tokens” (words or sub-word components…
11
Sycophancy is the first LLM "dark pattern"
People have been making fun of OpenAI models for being overly sycophantic for months now. I even wrote a post advising users to pretend that their work was…
12
Should LLMs just treat text content as an image?
Several days ago, DeepSeek released a new OCR paper. OCR, or “optical character recognition”, is the process of converting an image of text — say, a scanned…
13
LLMs are not superhuman at reading and writing neural network weights
A recent paper argues that frontier LLMs can work with neural network weights directly. This seems superhuman, but it isn’t. The paper shows that GPT-6 Astra…
14
Text AI watermarks will always be trivial to remove
The European Union AI Act will begin to be enforceable in August 2026, one month from now. One of the biggest new requirements is Article 50, which requires all…
15
Three kinds of AI products work
The very first LLM-based product, ChatGPT, was just the ability to talk with the model itself: in other words, a pure chatbot. This is still the most popular…
16
The first big AI disaster is yet to happen
The first public passenger locomotive, Locomotion No. 1, began service in September 1825. The first mass-casualty railway disaster happened seventeen years…
17
Crypto grifters are recruiting open-source AI developers
Two recently-hyped developments in AI engineering have been Geoff Huntley’s “Ralph Wiggum loop” and Steve Yegge’s “Gas Town”. Huntley and Yegge are both…