a_c

I’m using Jev to classify a blob of text I see in browser with an extension. Calibrated? No. But it’s handy enough. If a large blob of text is likely AI generated, I’m very likely to skip it

Thank you for all the explanations. I think I have better understanding now. Another question, at what scale (in terms of number of services, and maybe throughput? what other dimensions?) should we start considering using temporal?

Can I understand it as a performant celery with a UI?

Can someone shed some lights on what temporal does? Their blog says [1]

Durable Execution offers three key benefits:

    It improves application reliability by providing fault tolerance.
    It simplifies code by allowing it to focus on the goal instead of potential problems.
    It accelerates development by eliminating the need to write complex error-handling logic.

So... like exception handling? or something like erlang's let it crash mantra?

Why is it such a big deal? Genuine question, not trying to be snarky

[1] https://temporal.io/blog/what-is-durable-execution

I only started playing less than a year ago. I paid for one year because of game review to improve my game. Didn't know lichess exist, but I'm not paying for a second year for sure

Hiring for specific skills

Tax and regulations always punish the good actors

I feel that physical change the moment I saw my son first time. Something snapped/connected in my brain. I will do anything for that little thing. The feeling is visceral. It is very hard not to preach your best experience to others

Hey thanks for checking it out! Really glad the site has more audience than me and my son only and appreciate the feedback!

Mind if I ask which mode were you playing? Was it "Find a country" or "Guess a country"? It might be due to the countries being too small. I maintain a separate small countries list [1] from the bigger ones [2]

[1] https://github.com/kmcheung12/geo-dude/blob/main/public/data...

[2] https://github.com/kmcheung12/geo-dude/blob/main/public/data...

I didn't come up with a good way to show small countries so it is the way it is. Would love to hear suggestion though!

If you read this message, can you provide more context please? Are you choosing the "Name the country" mode with question from "Same continent". It might be that south america has only 13 countries in my database, so it appears to be repeated.

I just turned the repo public for your reference.

Country list at https://github.com/kmcheung12/geo-dude/blob/main/server/coun...

Picking a question from a continent is shuffle and then pick question per round, so "pick without replacement" https://github.com/kmcheung12/geo-dude/blob/main/server/inde...

Disclaimer, used a lot of AI so I didn't read every line of code

If you are not mindful, the tests it generates is just a re-implementation. Nothing was tested but every future change needs to modify two places

Thanks for the feedback, will look into it!

Nicely made! I thought I'm good with flags, and then there come a question of jordan, palestine and two more similar flags.

I made a similar but uglier little party game https://geo-dude.com for my kids. Supports up to 4 players

I have a similar project, with goodreaads records https://github.com/kmcheung12/goodshelf

I wrote the faceblurring. It didn’t work on multiple faces when I first processed the videos. Thanks for pointing out, I should have blurred them

Same!

Solid feedback, thanks!

1. is very doable. My local gym uses an app to manage monthly leaderboard already and people seem to like it. There is a friendly rivalry in my local gym where people "fight" for the top spots

2. A lot can be done from the helping beginner angle. A coach told me he wishes to show hip placement and movement to young climbers, as it is easier to show than to tell an 8yo. I thought of doing inverse kinematics from holds to resolve body positions, so I get beta (sequence of moves) for each problem right away. Too hard, don't know if it is solvable.

A bit tangential, I think in theory the climbing wall with a set of problems is itself an identifiable code. We might get rid of the QR code too. I can create an index of visual features that resolves to a camera position, raycast from camera position to narrow down to maybe 3-5 problems, and select videos from there. But I digress.

On route rotation, maybe the gym will be responsible for uploading the scan. My local gym uses an app to manage route. People view what route is available on a wall. It is just a label saying "red is a V4", no photos, and every month there is a leaderboard. People apparently use it. And occasionally people share their beta there.

For outdoor problems, I imagine it will be more like a map, and recording be taken with a drone. This is in contrast to the climbing magazine with plain photos of route.

I also think the usefulness decline as climbing ability increase. At the end of the day, climber just want to climb. But maybe there is some strava angle I can explore.

Appreciate your feedback, thank you!

Hey, looks like we both have a thing with 3D, love it! Mine is about bouldering, check it out https://news.ycombinator.com/item?id=49234599

Thank you!

Climbing gym photogrammetry. My idea is that route and climb recordings is best recorded against a 3D climbing gym model, rather than a plan video, and technology has gotten good enough to do with a regular phone.

As a climber myself, to climb effectively, one has to review their own climb, and with other people. 4D video, body positions across time, is very helpful in improving climbing. I managed to get a reasonable result (far from good) from one video, rather requiring multiple camera angles.

Haven't gotten to monetization yet, but I am enjoying every moment of making it.

A read only preview at https://kmcheung12.github.io/climb-preview/w/ae43c6e5

Tangential, for static site like this, why do people choose to host on vercel instead of github page?

Writing code with real use is very different from writing code for the sake of writing (looks cool, sounds cool).

Writing code with one user is the base case of a “why am I writing this” recursion - because I use it.

Every action leads to next insight. For me, a browser extension that provides a table of content to chat interfaces because I always need to scroll up and down to read preview queries. Writing browser extension leads me to writing a RSS reader navigating with only keyboard. Writing RSS reader leads me to writing a virtual book shelf that visualize Goodreads record on a virtual 3D bookshelf. Writing virtual book shelf lead me to making a virtual climbing gym where I scan my local climbing gym into 3D model where I can map climbing videos on it. And it is just one chain

Accurate reconstruction is entirely up to the camera's ability to orbit around the object. If we are talking about cracks, the visual signal will be very poor as we can't wiggle a camera inside a crack.

On the other hand, if we want a 3D reconstruction of a route in general, using a drone, then that's very doable. We still rely on 2D climbing magazine/book/app to view a route, where the contour of the route itself is highly desirably imo.

I scanned my local climbing gym and have climbing videos mapped onto the wall

https://kmcheung12.github.io/climb-preview/w/a6ece004

I think 3D application is massively under utilised. There is whole family of interaction the textual web cannot provide

They are going to reset the weekly limit before your weekly limit. Mine resets tomorrow too.

My experience from running a django app with thousands of migrations and dozen of apps

- Testing involving models is slow. It runs all your migration. Python's own "Unittest" is fast for functions not involving models.

- The suggested pattern for business logic is "active record", which is putting functions about a model alongside the model itself. Good for clean case. Doesn't really work when your operation involve multiple models.

- But if you put function about a model inside a model, the migration doesn't serialize the function into migration history (I could be wrong)

- "Data" migration is handy when you want enum-like data in database available to all teammates. e.g. list of currencies

- It is very tempting for new team member just to create an django app, with its own view and model, because it feels like starting a fresh without needing to care about existing business context. On the other hand, designing conceptual boundary between apps is very important. Because starting app is easy, and often (not always) wrong.

- The squash migration utils are limited because I guess of relatively low usage. It squashes linear migration history (with manual curation of starting and ending migration node I think? Not sure about latest state). I wrote a util to squash whole migration history by finding linear segments and squash those segment of history one by one. Didn't release it, but wrote some notes https://gist.github.com/kmcheung12/08b4c234b38c23efd43550da9...

Is there any write up regarding the SimdJson Algo? Definitely love to read more of it!

People learn and relearn the core concepts from different paths. A recent trend due to coding agent is that people are learning software engineering is not about syntax and frameworks. It is just the way it is. People build up concepts with things they already understand. The same happens when someone doesn't use a book. It just went implicit and way harder for you to find the magic combo.

I didn't update it for a while. It reads your local conversation history, correlate input/output token with usage (claude doesn't give exact breakdown), and classify if they are subagent/tooluse/normal message/whatever. Let me know if I can make it any way useful to you.

https://github.com/kmcheung12/ccaudit