Check this out: #CDK construct to manage route53 without zone delegation legacy
constructs.dev/packages/@ca...
Using it for all for all domains, subdomains, environments and multi-account app deployments. Deployed it and never had to touch #Route53 zones ever again.
Getting a new dev laptop!
#Coder456DevSetup
#FrontEndMan
#GetAdminAccessGame
#UpdatePasswordGame
#EnterPin3TimesGame
#CongratulationsOnLaptopSetup
#RedGreenDeployment
#MarblePairProgramming
#SinglePaneOfGlass
#FinalGitPush
#NotAllowedToStopCoding
AI coding agent getting lost in 1200 lines class with pretty good doc strings.
40 min wasted, could've coded it 80% already🥹
ChatGPT tries to design DynamoDB 😱
It said Ddb will be slow, inefficient, adds jitter, saturation, and cost 😂
Yes it’s designed to be additive. Lots of people getting hooked
Looks like a lot of shell scripting, by the way a good and robust way to unit test shell scripts is shpy
github.com/codehearts/s...
✨ In March 2023, I published an article about AI.
This was right after OpenAI showcased how GPT4 could turn a hand-drawn sketch of a website into real HTML/CSS/JS. The consensus online was that FE jobs would stop existing within 1-2 years.
Well, it’s been 2 years! So, I just published a follow-up:
interested in this piece. What kind of trans? Could you elaborate how it felt?
Thats like 600$/mo
What is it making?
Would you like to share detailed experience? I’m collecting experiences for a post on ai coding addiction
How I feel vide coding
#VibeCoding
AWS Amplify Gen2 has a fix for the unreadable CloudWatch log names
Now you get:
/aws/lambda/astR
instead of
/aws/lambda/amplify-fw34ufwj3m-master-b-astRlambdaD70DF3-YDtuB774
Quick demo in the video
PS: No env info or prefixes in that name, but it's easy to add
Some of the approaches you listed would get developers into lots of trouble: splitting stacks, extending Stack construct to define resources
The tests runtime in #CDK was pretty bad
I looked how to fix it and #Typescript and #Jest had options to do it better. But its complicated.
thecloudtimes.substack.com/p/strategy-f...
The personal style and interactivity is what makes it great.
My interactive AWS NAT Gateway blog post is now published. Check it out at malithr.com/aws/natgatew....
Looks so polished!
Did it take a lot to build?
bsky.app/profile/sill...
It is also unbelievably slow at getting anything over the finish line
bsky.app/profile/alex...
Doesn't look magical at all when looking at the expected time to generate some lines of code
bsky.app/profile/jiti...
Vibe Coder Looking AT Time Distribution of Expected Time to Generate 100 LOC
bsky.app/profile/gene...
Trying is nice but measuring is even better. Show your boss some numbers!
bsky.app/profile/ashb...
It "speeds up" only the test of the developer's patience...
bsky.app/profile/fais...
The new thing doesn't look particularly accelerating you know
With an expected rate of 100 LOC every 2 hours...
bsky.app/profile/notn...
Let's be kind to CS graduates, they'll be fixing all the “vibe coding”.
I propose brute-force coding!
bsky.app/profile/kout...
For one engineer the Monte Carlo simulation shows how painfully inefficient and slow code "vibrating" is. Can imagine any dev team would tolerate this.
bsky.app/profile/emil...
Yup, developers don't brute-force
Man, this whole AI coding thing is giving some deep, deep hippo-doo-doo vibes...
github.com/OperationalF...
#AICoding #VibeRating #MonteCarlo
Further confirmed by actually coding simulation with AI
Took one for the team: 8 hours for 400 LOC
Just as predicted. YMMV, plug in what you think realistic parameters are and run for yourself.
#AICoding #VibeRating #MonteCarlo
How long does it take for AI coding agents to do things? I mean produce valuable thing like piece of code that works. Rather than output, which is cheap.
I wrote a script (with an AI agent, of course) for the Monte Carlo simulation. With realistic params, it shows there is a... problem
Left Cursor opened and it hammered disk non-stop at 15 MB/s rate
