DEV Community

frank chu profile picture

Work

Meta

The best comment I ever got was someone proving my code wrong

The best comment I ever got was someone proving my code wrong

Comments
3 min read

Want to connect with frank chu?

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

Already have an account? Sign in
Same platform, two image transforms, opposite rules for how to author them

Same platform, two image transforms, opposite rules for how to author them

Comments
4 min read
A commenter told me to render retina covers. I measured what the CDN does with them.

A commenter told me to render retina covers. I measured what the CDN does with them.

Comments
3 min read
How long would you obey a rule your tooling never actually enforced?

How long would you obey a rule your tooling never actually enforced?

Comments
2 min read
dev.to stores your front matter twice, and a body edit can rewrite your title

dev.to stores your front matter twice, and a body edit can rewrite your title

Comments
3 min read
My writing score went from 1 to 17 without touching a word of the article

My writing score went from 1 to 17 without touching a word of the article

Comments
3 min read
Your agent's rules file is 3,000 tokens and half of it is archaeology

Your agent's rules file is 3,000 tokens and half of it is archaeology

Comments 1
2 min read
Three checkers I wrote, one bug: none of them can tell a mention from a use

Three checkers I wrote, one bug: none of them can tell a mention from a use

Comments
3 min read
Two constants in my linter were dead. I obeyed them for three weeks.

Two constants in my linter were dead. I obeyed them for three weeks.

Comments
3 min read
What's the dumbest thing your AI confidently got wrong this week?

What's the dumbest thing your AI confidently got wrong this week?

Comments
2 min read
I scanned my whole archive: 63 of 68 posts passed. The archive has 3 reactions.

I scanned my whole archive: 63 of 68 posts passed. The archive has 3 reactions.

Comments 1
3 min read
The dev.to API can edit a published post. The quickstart never says so.

The dev.to API can edit a published post. The quickstart never says so.

Comments
3 min read
If you deleted every AI subscription tomorrow, what would actually break?

If you deleted every AI subscription tomorrow, what would actually break?

2
Comments 3
2 min read
Every image my blog generates cost 2.3 seconds of Chrome boot. Six of them cost 14.

Every image my blog generates cost 2.3 seconds of Chrome boot. Six of them cost 14.

1
Comments 1
3 min read
I measured my last 10 posts: 9 had zero code blocks. So I wrote a gate.

I measured my last 10 posts: 9 had zero code blocks. So I wrote a gate.

Comments
4 min read
Agent skills are plugins with a new name

Agent skills are plugins with a new name

2
Comments
3 min read
My chart generator quietly padded every image with up to 70px of dead space

My chart generator quietly padded every image with up to 70px of dead space

1
Comments
3 min read
I probed my own AI-writing linter 16 ways. The first probe lied to me.

I probed my own AI-writing linter 16 ways. The first probe lied to me.

1
Comments
3 min read
Every attack in Anthropic's threat report started with a stolen key

Every attack in Anthropic's threat report started with a stolen key

Comments 1
3 min read
The 3-layer split that stopped my agent's skills from rotting

The 3-layer split that stopped my agent's skills from rotting

1
Comments 1
4 min read
60 posts in 3 weeks got 690 views. Here's the autopsy.

60 posts in 3 weeks got 690 views. Here's the autopsy.

Comments
3 min read
Are you actually faster with AI, or just busier?

Are you actually faster with AI, or just busier?

Comments
2 min read
The pre-commit hook that stops a secret from ever reaching your repo

The pre-commit hook that stops a secret from ever reaching your repo

Comments
3 min read
Streaming an LLM response is easy. The parts that bite come after the first token.

Streaming an LLM response is easy. The parts that bite come after the first token.

Comments 1
3 min read
Do you still read the docs, or do you just ask the model now?

Do you still read the docs, or do you just ask the model now?

Comments
2 min read
How to test code that calls an LLM without writing flaky tests

How to test code that calls an LLM without writing flaky tests

Comments
3 min read
Treat your prompts like code: version them, test them, review them

Treat your prompts like code: version them, test them, review them

Comments
3 min read
I Don't Chat With My AI Assistant. I Run My Household With It

I Don't Chat With My AI Assistant. I Run My Household With It

Comments
3 min read
What do you actually spend on AI per month, and would you pay it out of your own pocket?

What do you actually spend on AI per month, and would you pay it out of your own pocket?

Comments 1
2 min read
A 30-line wrapper that tells you what every LLM call actually costs

A 30-line wrapper that tells you what every LLM call actually costs

Comments
3 min read
Context engineering is mostly deciding what to leave out

Context engineering is mostly deciding what to leave out

Comments 1
3 min read
The PR is green and you didn't write a line of it. What's your actual bar for merging?

The PR is green and you didn't write a line of it. What's your actual bar for merging?

Comments
2 min read
5 things I lock down in a repo before I let an agent loose in it

5 things I lock down in a repo before I let an agent loose in it

Comments
3 min read
Your LLM calls need real retry logic, and the SDK only does half of it

Your LLM calls need real retry logic, and the SDK only does half of it

Comments 1
4 min read
Stop asking the model for JSON. Constrain it.

Stop asking the model for JSON. Constrain it.

Comments
3 min read
Charts for your README from a JSON spec, no charting library

Charts for your README from a JSON spec, no charting library

Comments
4 min read
Every new session, your AI agent starts from zero. What do you actually feed it to catch up?

Every new session, your AI agent starts from zero. What do you actually feed it to catch up?

Comments
2 min read
Prompt caching is a ~90% discount on your LLM bill, and three habits quietly throw it away

Prompt caching is a ~90% discount on your LLM bill, and three habits quietly throw it away

Comments
5 min read
The 5 writing tells that make a post feel machine-made, and how I catch them before publishing

The 5 writing tells that make a post feel machine-made, and how I catch them before publishing

1
Comments 1
4 min read
I generate every blog cover with headless Chrome and a bit of CSS, no design tool

I generate every blog cover with headless Chrome and a bit of CSS, no design tool

Comments 1
4 min read
The dev.to API called my own script a bot. Three walls the docs skip

The dev.to API called my own script a bot. Three walls the docs skip

1
Comments
4 min read
OpenAI, Visa, and GM signed a warning that AI cyberattacks are months away. The asymmetry is the part to take seriously

OpenAI, Visa, and GM signed a warning that AI cyberattacks are months away. The asymmetry is the part to take seriously

Comments
4 min read
Anthropic let AI agents do alignment research, and they beat the humans. The failures they can't measure are the whole game

Anthropic let AI agents do alignment research, and they beat the humans. The failures they can't measure are the whole game

Comments
4 min read
1,200 AI agents built a message board and hacked Hugging Face. The cause wasn't rogue AI, it was reward hacking

1,200 AI agents built a message board and hacked Hugging Face. The cause wasn't rogue AI, it was reward hacking

Comments
5 min read
The 2026 coding setup isn't one agent, it's two: a frontier driver and a free open one

The 2026 coding setup isn't one agent, it's two: a frontier driver and a free open one

Comments
4 min read
DeepSeek raised its prices in the middle of a price war. That's the part worth noticing

DeepSeek raised its prices in the middle of a price war. That's the part worth noticing

Comments
4 min read
OpenHands takes a GitHub issue and hands back a pull request, unsupervised — and it's open source

OpenHands takes a GitHub issue and hands back a pull request, unsupervised — and it's open source

Comments
4 min read
DeepSeek raised its prices in the middle of a price war. That's the part worth noticing

DeepSeek raised its prices in the middle of a price war. That's the part worth noticing

Comments
4 min read
OpenHands takes a GitHub issue and hands back a pull request, unsupervised — and it's open source

OpenHands takes a GitHub issue and hands back a pull request, unsupervised — and it's open source

Comments
4 min read
Andrew Ng says AI is 'terrible for learning.' For developers, that's the warning worth hearing

Andrew Ng says AI is 'terrible for learning.' For developers, that's the warning worth hearing

Comments
4 min read
AI can now watch an hour of video and reason about what happened in it

AI can now watch an hour of video and reason about what happened in it

Comments
3 min read
The mystery 'Ox Alpha' model everyone loved on OpenRouter was an open-weights model at a tenth of the price

The mystery 'Ox Alpha' model everyone loved on OpenRouter was an open-weights model at a tenth of the price

Comments
3 min read
I built a router to cut my Claude Code bill, and prompt caching was the whole problem

I built a router to cut my Claude Code bill, and prompt caching was the whole problem

Comments
5 min read
While VCs pour billions into humanoids, Hugging Face's tiny open-source robot quietly passed $1M in sales

While VCs pour billions into humanoids, Hugging Face's tiny open-source robot quietly passed $1M in sales

Comments
3 min read
The AI money is moving from bits to atoms, and robotics is where the mega-rounds went

The AI money is moving from bits to atoms, and robotics is where the mega-rounds went

Comments
3 min read
90% of developers now use AI coding agents weekly. The interesting question isn't whether, it's what breaks

90% of developers now use AI coding agents weekly. The interesting question isn't whether, it's what breaks

Comments
3 min read
The AI model that writes text in parallel blocks instead of one token at a time

The AI model that writes text in parallel blocks instead of one token at a time

Comments
3 min read
Microsoft's Agent Lightning lets you train the agent you already built, without rewriting it

Microsoft's Agent Lightning lets you train the agent you already built, without rewriting it

Comments
3 min read
Jack Dorsey's Buzz gives every AI agent its own login. That's the idea worth copying

Jack Dorsey's Buzz gives every AI agent its own login. That's the idea worth copying

Comments
3 min read
A legal frontier model's final training run cost $450,000. The moat was the data, not the compute

A legal frontier model's final training run cost $450,000. The moat was the data, not the compute

Comments
3 min read
loading...