< Back to the archive

Like what you see? Subscribe here and get it every week in your inbox!

Issue #284 - August 18, 2024

Here are the top threads of the week, happy reading!

Top comment by oldprogrammer2

Stripe, Block, and PayPal each solved a massive pain point.

PayPal provided a way to pay people and vendors without giving away your credit card number.

Square made it easy to accept payment in person on a phone, without an extensive upfront underwriting experience and without expensive fixed monthly fees.

Stripe did the same as Square, but for accepting online payments.

Fraud and Risk come in many forms, and these providers, even with their UX innovations, sit on top of those same rails to reduce fraud. Without those rails, buyers can’t trust sellers and sellers can’t trust buyers.

In my opinion, you need to find a way to solve that problem before you can eliminate the fees being captured by these providers.

Top comment by lolinder

In my own experience and watching similar things play out with friends and family, compulsive gaming of the kind you're describing (assuming you're not exaggerating) is almost always covering for some kind of stress or anxiety. More specifically, the most common cases where I've seen this kind of behavior have been with kids whose parents were overprotective and constantly managing their business while they were growing up.

I'm going to be brutally honest here because you seem to be open to it: Kids who reach adulthood without any practice adulting tend to become overwhelmed by the sudden onset of real responsibility, and it's not uncommon for that to lead to maladaptive coping mechanisms, with video games being one of the most benign. In this context, since you're asking and clearly still extremely involved in your adult child's life, I strongly suspect that this is exactly what's going on here (though obviously I could be wrong!).

If I'm right, managing their life for them and trying to solve their problems with university will provide a band-aid, but won't address the root of the problem. You might get them through uni and have them fall apart when they get into a career or a relationship. At this point you have to let go—it's too late to make the transition to adulthood smooth, but better a bumpy transition at 19 than complete failure to launch.

Top comment by kevg123

> What are the key assets you monitor beyond the basics like CPU, RAM, and disk usage?

* Network is another basic that should be there

* Average disk service time

* Memory is tricky (even MemAvailable can miss important anonymous memory pageouts with a mistuned vm.swappiness), so also monitor swap page out rates

* TCP retransmits as a warning sign of network/hardware issues

* UDP & TCP connection counts by state (for TCP: established, time_wait, etc.) broken down by incoming and outgoing

* Per-CPU utilization

* Rates of operating system warnings and errors in the kernel log

* Application average/max response time

* Application throughput (both total and broken down by the error rate, e.g. HTTP response code >= 400)

* Application thread pool utilization

* Rates of application warnings and errors in the application log

* Application up/down with heartbeat

* Per-application & per-thread CPU utilization

* Periodic on-CPU sampling for a bit of time and then flame graph that

* DNS lookup response times/errors

> Do you also keep tabs on network performance, processes, services, or other metrics?

Per-process and over time, yes, which are useful for post-mortem analysis

Top comment by tazu

If you have access to target customers (hopefully you do), I'd recommend a short questionnaire using Van Westendorp's Price Sensitivity Meter [1]. I've had success using it in the past, and you only have to ask four questions:

1. At what price would you consider the product too expensive?

2. At what price would you consider the product too cheap (low quality)?

3. At what price would you consider the product expensive enough to make you nervous?

4. At what price would you consider the product to be a bargain?

You can then graph the results and pick the optimal price (for starting, at least).

[1]: https://en.wikipedia.org/wiki/Van_Westendorp%27s_Price_Sensi...

Top comment by keiferski

It’s just the typical aesthetic model used and isn’t inherent to the tech itself. It’s very easy to make AI images in specific art styles, with the result that you can’t tell they’re not real.

This is actually something of a pet peeve of mine - people sharing AI images never use styles other than the generic shiny one, and so places like Reddit.com/r/midjourney are filled with the same exact style of images.

Edit: if you’re looking for other style inspiration ideas, this website is a great resource for Midjourney keywords: https://midlibrary.io/styles

Top comment by mattnewton

While not quite your definition of toy, I have a small deep learning rig I built myself with two 4090's, and that has been enough to train several different ~200m parameter LLMs, starting with a hand rolled tokenizer and just vanilla pytorch to experiment with different architectures. While its not going to win any benchmarks or be usable for real problems (you should just be fine tuning llama), it has been super valuable for me to really understand exactly how these things work.

I use devpod.sh and a pytorch dev container I can spin up locally, with the intention of also spinning it up in the cloud to scale experiments up (but I haven't done much of that). Still, can recommend devpods for reproducible environment I don't feel worried about trashing!

If people are interested I can throw the git repo up now, but I have been planning on finding some time clean it up and write up a really short digest of what I learned.

Above anything I can write though, I highly recommend Andre Kaparthy's youtube channel - https://www.youtube.com/@AndrejKarpathy You can follow along in a google colab so all you really need is a web browser. My project started as following along there and then grew when I wanted to train it to mimic my friends and I on some data I had of us chatting in slack, which meant some architecture improvements, figuring out how to pre-training on a large corpus, etc etc

Top comment by whalesalad

Couple weeks ago we had an issue where we were running into the 100 MB file upload limit so we upgraded to pro. Even after upgrading, we still are stuck at 100 MB upload limit and after weeks of trying to reach out to support (submitting HAR files, etc) that limit has not been raised. Cancelled pro because it simply doesn’t work and yeah support sucks. At one point we got a candid response (paraphrasing): sorry we've been late to reply, shit has been hitting the fan recently.

It's really unfortunate because Cloudflare is incredibly valuable tech and we rely on it for all sorts of things: zero trust, tunnels via cloudflared, advanced page/routing/redirect rules, a lot of cloudflare pages/workers deployed, on top of all our DNS there.

We manage about 800k domains so having a reliable partner here is critical.

Top comment by blntechie

I love the convenience of physical SIMs. Will hate the day when I don't have a choice but to 'upgrade' to eSIMs.

I have a phone with dual physical SIM card slots. I can go to any country in the world, buy a SIM, put it in, and am up and running. eSIM provisioning at airports is barely available in few coubtries.

I broke the display of my phone? Easy, remove the SIM and put it in a spare phone and I can still make and receive calls.

Top comment by lambdaba

You shouldn't dismiss Lua outright, it's a language designed specifically for embedding, there's practically nothing to learn to begin using it productively.

See this presentation by a core Neovim maintainer that explains why Lua was chosen https://www.youtube.com/watch?v=IP3J56sKtn0.

See also http://lua-users.org/wiki/MechanismNotPolicy to understand Lua's core design.

Top comment by JohnMakin

One of the funniest things I've ever heard/read about chatGPT "writing" is that it writes like a highschool student trying to inflate their word count on an essay. It typically uses a lot of words to say very little, and the style is hard to un-see once you recognize it.