This is awesome and I love it. Thanks for sharing! I'm just wondering what made you wanna be a pilot instead of sticking with a regular software engineering career, especially since you’ve got a degree in that and, of course, your talent here :)
HN user
Crier1002
thanks! apologies i was on mobile and missed this. im excited to try it out
ive always wanted this. im going to give it a go!
does this happen to support IDE like vscode?
i second this! datasette is amazing and i've been using it to publish data that i collect and find interesting
do you have examples i can refer to? i’d love to learn to write better docs
i found Soft Skills Engineering Podcast this year and i'm enjoying it!
over the past few years: Dune Part I/II. I love the story telling, the visuals, the character building, the intensity etc.
mermaidjs is my default go-to. it works super well with Claude 3.5 if you ask it to visualize <something> in mermaidjs syntax
i experienced this too! it happened to my work account but not on my regular personal account. i’m not sure what happened there
this sounds like a cool story to write about!
i built a chrome extension recently and i was evaluating/trying out different frameworks for building a browser extension.
for anyone who is interested for an alternative to Plasmo, i chanced upon https://github.com/extension-js/extension.js
Mermaidjs is amazing. you could ask LLM to visualise your question (e.g codebase) using mermaidjs syntax and it mostly works just fine without any syntax errors. or at least on Claude Sonnet 3.5 from my experience
i often browse the web for fun (but in a careful way), I can totally see how changing up the CSS class names in the site’s HTML regularly would mess up a bunch of the XPATH/CSS selectors in the crawler. It’d seriously be a nightmare for me if the site owners could just flip a 'switch' and change the class names easily
I use https://github.com/litl/backoff for configurable backoff + retry
context: OpenAI API used to be super flaky back in the early days, i needed to retry my requests quite frequently and i found this
this! im selling a subscription with this pricing model and it’s quite painful to swallow. having that said i really enjoy using lemon as my payment processor
"work smart, not hard"
i think this advice is a bit too vague and it's just something people say to sound smart to trick people into thinking there's a shortcut to mastery
i genuinely believe that if you work hard to hone your skill on something, it will do you good. if you want to get good at something, you gotta put in the hours, the grind. e.g. to learn how to play a tennis, you gotta, you know, actually get down to the court and play tennis instead of watching tutorial on how to play tennis. Same goes with programming, music, etc.
now pair working hard with the ability to retrospect -- you'll master a lot of skills in life
just wanted to say actively trying to get bored is some of the best things that i've done for myself as an adult in a dopamine-driven world (exercising aside). i'm not saying to do nothing at all, but trying to be bore yourself whenever you're off work. some examples are like instead of scrolling your phone whenever you're in public transit, just do nothing.
i hope they'll eventually support Mermaid (https://github.com/mermaid-js/mermaid) for creating diagrams directly within documents. i've been using it a lot for my markdown files and it works amazingly well with LLMs (e.g. asking LLM to generate the diagram representation of something using Mermaidjs)
what an amazing read!
slightly unrelated to the post: does anyone know any similar blog posts as such about people writing about their life? especially in areas of the world that is less known of?
another one that i really like is https://brr.fyi/ where he posts life working at the antarctica
out of curiosity: is there a specific reason to use robinraju/release-downloader@v1 over actions/download-artifact@v4 here at your 'Download previous DB' step in build_db.yml?
your comment immediately reminded me of "We’re all just temporarily abled" (https://news.ycombinator.com/item?id=37205731)
check out https://github.com/mlc-ai/web-llm
IMO the main limitation is access to powerful GPUs for running models locally and the size of some models causing UX problems with cold starts
No. it's perfect as what is it already.
standing desk and a monitor for sure.
i think it help to expedite learning/research by quite a bit. before this, i had to append "site:reddit.com" to my google searches and spend hours to read multiple post/articles about the topic.
now, i do my learnings/research from something like phind or perplexity. I have a shortcut "!pl" or "!p" setup on my address bar. I just have to type in "!pl food recommendation for keto diet" for example in my address search bar and it would summarize everything for me. After that, all i had to do is to read/click a few more links in deep into reddit or wtv to verify what the LLM tells me is according to it's citations, then i can gauge if the answers are credible or not. i just ask some follow-ups if i have any. The result has been quite satisfactory so far
when i make something and then share it with the world and people actually like/use (sometimes even pay for it)
https://www.reddit.com/r/EngineeringResumes/wiki/index/ has some pretty good guide/examples
I am leaning towards LC style interviews over a take home assignment that is meant to take “no longer than 2 hours” to complete
ive always enjoyed reading about your inspiring TinyPilot journey over the past few years. i hope you won’t stop writing! congrats and all the best!
https://github.com/mlc-ai/web-llm basically bringing the power of llms to web browsers via WebGPU. i think it’s useful because it’s better for privacy as nothing leaves your browser.
having that said i dont see how this is feasible yet on a widespread public scale given how computational intensive this is.