ivanstojic

> If you asked me six months ago what I thought of generative AI, I would have said

It’s always this tired argument. “But it’s so much better than six months ago, if you aren’t using it today you are just missing out.”

I’m tired of the hype boss.

When tractors were invented, there was a notable reduction in human employment in agriculture in the USA. From a research paper (https://faculty.econ.ucdavis.edu/faculty/alolmstead/Recent_P...):

> The lower-bound estimate represents 18 percent of the total reduction in man-hours in U.S. agriculture between 1944 and 1959; the upper-bound estimate, 27 percent

I'm not seeing that with LLMs.

At the same time, I actually wouldn’t mind a world in which AI agents cost $5000 a month if that’s what companies want to charge.

I feel like at some level that would remove the possibility of making a “just as good as humans but basically free” arguments and move discussion in the direction that feels more productive: discussing real benefits and shortcomings of both. Eg, loss of context with agents vs HR costs with humans, etc…

I started reading the article and immediately got hit by the incorrect statement in the opening:

> If AI agents help each support employee handle 30% more tickets, that's like adding 30 new hires to a 100-person team, without the cost.

I think this is an oversimplification designed to make LLMs seem more profitable than they actually are.

TL;DR: on-call manages acute issues, documents steps taken, possibly farms out immediate work to subject matter experts. Rate on-call based on traces they leave behind. Separate on-call with same population, but longer rotation window handles fixes. Rate this rotation based on root cause reoccurrence and general ticket stats trendlines.

Longer reply:

I have on-call experience for major services (DynamoDB front door, CosmosDB storage, OCI LoadBalancer). Seen a lot of different philosophies. My take:

1. on-call should document their work step by step in tickets and make changes to operational docs as they go: a ticket that just has "manual intervention, resolved" after 3 hours is useless; documenting what's happening is actually your main job; if needed, work to analyze/resolve acute issues can be farmed out

2. on-call is the bus driver, shouldn't be tasked with handling long term fixes (or any other tasks beyond being on-call)

3. handover between on-calls is very important, prevents accidentally dropping the ball on resolving longer time horizon issues; handover meetings

Probably the most controversial one: separate rotation (with a longer window - eg. 2 week) should handle tasks that are RCA related or drive fixes to prevent reoccurrence

Managers should not be first tier on any pager rotation, if you wouldn't approve pull requests, you shouldn't be on the rotation (other than as a second tier escalation). Reverse should also hold: if you have the privilege to bless PRs, you should take your turn in the hot seat.

Thanks for being here over the years. The faces changed some, stayed the same some, but it’s good people and good conversation!

May you all have a happy new year, and many more!

Tossing mine in the pot too: make + pandoc: https://ordecon.com/2020/07/pandoc-as-a-site-generator/index...

I've been playing around with a limited PCB autorouter for mechanical keyboards. You feed it a KLE layout file, and it spits out a KiCad PCB, layout map for QMK firmware, and various SVG cuts for case machining.

The country lines on the globe are ridiculously out of date, and show countries that haven’t existed for 30 years.

How would this work against scrapers that are based on driven anpproved browser instances, eg. something like Selenium?

The length of a law doesn’t reduce its murkiness - in fact it makes it more pronounced. That is why (in the common law systems) there is so much discussing and re-discussing of topics that plenty of other cases already covered, using new approach angles. If you make laws longer and more complex, you only make them serve more those that have the budget to explore all branches of the decision tree.

What happens with eSIM devices?

I had to research who that is. I stumbled onto this:

https://www.tabletmag.com/sections/news/articles/red-pill-pr...

Loaded associations.

My TTRPG group has been playing together for several years - on and off. We'd been mostly playing D&D 5e, even though we had members clamoring for other systems. After the WotC fiasco started blowing up, we finally bought a couple of Pathfinder 2e books, ran a few trial characters and one-shots. We are now in process of starting a new campaign entirely based on Pathfinder that we expect could take a couple of months to play out.

It really only took us one session to figure out the major mechanical systems in the gameplay, and a few we are still learning about. The major component remains the role playing experience which is the same, no matter which dice you throw to decide outcomes of situations.

I think Hasbro will soon realize that while the social impact of the name D&D is deep, the stickiness of the system is much lower than they gambled on. It's unlikely that any of my member group will soon buy anything new from them.

This is true, but only by omission: we don't know everything he believes. We do know he publicly supports the genocide in order to protect himself. MLK day is around the corner. It's worth considering what it cost some people to do the right thing, vs what some people got paid to support the wrong thing.

Seeing stories like this makes me very averse to the idea of running a business myself.

I don’t mind the idea of trying and failing because the market isn’t there or my execution sucks, but trying and having my earnings be trapped in an algorithmic black hole with no customer support - no thank you. I don’t think there’s any other kind of business interaction that works in this way.

Eg, I can’t imagine one day waking up to my electric company unplugging my house and refusing to talk to me ever again.

How does this differ from XUL?

Isn't the real performance gained by eliminating the triple nested for loops and memoizing over a sliding window, rather than the use of XOR? You would get the same order-of-magnitude improvement if you used a map combined with iterate-only-once approach.

Ie, it's undeniable that the final program works much faster, but the article should really be called "a smarter memoization trick."

This is to me peak social media in 21st century. Somebody has been curating or authoring content for 13 years, but if they don't have clearly identifiable personalities that you can track down, we ask ourselves if they are real.

In general, I don't care about the people behind the article content. They have their own lives that should be private, and the articles should be able to stand on their own merit.

Kura Sushi (https://kurasushi.com/) that's mentioned in the comments is one place, but there are also a few others whose names I can't remember in the Sawtelle Japantown. It's worth noting that Kura is highly automated and you don't have to interact directly with staff after you are seated.

All of them basically center their feel around the futuristic automation spiel.

There’s a number of restaurants in LA that use robots for delivering drinks to your table. I found them strange the first time I saw them but eventually they become invisible.

I’ve never yet seen them misbehave in a notable way, but they seem like nothing more than trays with wheels and basic path finding.

Working on mechanical keyboard design. I learned so much about CNC and other manufacturing methods, relearned a lot about electronics and most entertainingly learned more than I ever thought I’d know about the USB protocol.

We should recognize that there are universal needs for geezers and kids: we want some form of personal security. For a lot, that means home ownership.

Let’s look at data: If the geezer is 60 and paid off their loan today, they got it 30 years ago (in 1992). The median price of single family freestanding homes in LA county in 1992 was 221k. Today it’s 860k. In 1992 the median household income was 32k and in 2020 (last available data point) it’s 76k.

I understand that “the geezers” have something to teach us with a perspective of a lifetime and the needs all humans share.

At the same time, we have to acknowledge and account for the economics having totally changed.

We easily claim the opposite for other companies (Apple, Facebook, Google) and other countries (mostly Asian fabrication centers). Why do we have such a strong belief that Uber USA != Uber Germany?

(Edit: I'm Croatian and fairly well aware of how taxis work in Europe)

After years of playing it and putting probably thousands of hours into Angband, I switched to Dungeon Crawl Stone Soup (https://crawl.develz.org/) and I love it to bits.

Text mode, turn based dungeon crawlers aren't for everyone, but I love them: they are very tactical, but don't feel like a typical tactics game, have a fun fantasy based lore, and are just a blast. And most importantly, you can pause and quit at any given moment, and continue just as easily.

> There IS risk to the employee; they now have a real loan outstanding and 100% personal recourse, so if the common stock becomes less than exercise price, their personal assets are on the hook

https://twitter.com/theryanking/status/1493390184897032201?s...

Does anyone else feel awkward about the use of the word "attack" in this context?

Interesting thing to remember: quaternions are a special application of rotors, a concept which may be easier to understand as a base:

https://marctenbosch.com/quaternions/

If you’ve never seen the “five whys” technique before, then you are probably unfamiliar with its criticisms as well.

This shows a broad spectrum of proponents and critics of the approach: https://www.google.com/search?q=five+whys+site%3Anews.ycombi...