@drewww@social.coop avatar

drewww

@drewww@social.coop

ex. research & data @twitch, phd @medialab, engineer @olincollege. he/him.

interested in: socio-technical systems, computer-mediated communication, online communities, data visualization

playing: card games, roguelikes, mobas, wargames, 18xx.

This profile is from a federated server and may be incomplete. Browse more on the original instance.

drewww, to random
@drewww@social.coop avatar

Undeniably these the new Dune adaptations are excellent. But I wonder if part of what's helping them land is the social media "explainer" ecosystem that supports the broader discourse around them. The '84 Lynch Dune had to more or less stand on its own as a media artifact. You could read the book, but that's a much higher bar than absorbing the barrage of 60s explainers I'm getting exposed to now.

drewww, to random
@drewww@social.coop avatar

I'm kicking off this morning! I'm going to build "runner" - a cyberpunk escape roguelike. I spent the last ~week re-learning javascript/typescript, setting an end-to-end build system, and assembling a basic "engine" on top of rot.js. I've got some basic models for UI elements, transitioning between screens, FOV, and a few other basic capabilities. I'll use this thread as a devlog for updates along the way!

A top down view of an environment rendered using single characters. The player is a yellow '@', the floor is '.', the exit is a green '>'.

drewww,
@drewww@social.coop avatar

Code lives here: https://github.com/drewww/runner
In progress builds will be available here: https://drewww.itch.io/runner

drewww,
@drewww@social.coop avatar

Day 1 goals:

  • make a debug level generator that has a lot more open space for testing
  • develop enemy behaviors (move forward until you it something, spinner-type immobile, maybe a "hunter" type)
  • develop enemy "sight" rules, i.e. they can see a long way in one direction.
  • make enemy "vision" lighting interact, i.e. multiply a space's color when people lasers hit it
  • [reach] start to work on player movement framework
drewww,
@drewww@social.coop avatar

Well that went faster than expected. Checklist done before lunch. I don't have any intuition about multiply versus adding in color; turned out add was the "better" effect for showing a space with multiple entities seeing it at once. After lunch -- movement.

drewww,
@drewww@social.coop avatar

Movement went well! I'm proud of solving some tricky design issues. Full reflection in the dev log on itch: https://itch.io/post/9495342

drewww,
@drewww@social.coop avatar

Woof, game design is hard. I've got a vertical slice working but it's not fun (yet?) More detailed discussion over on itch: https://itch.io/t/3560940/runner-a-cyberpunk-escape-roguelike#post-9508010

drewww,
@drewww@social.coop avatar

Gaze upon my terrible first level generator.

drewww,
@drewww@social.coop avatar

Day 4 complete. Cracked some of my ennui from yesterday with the addition of a major new antagonist: the hunter. It spawns "behind" you from where you entered the level, and you have to use your movement abilities to stay one step ahead of it.

Full dev log here: https://itch.io/post/9512278

drewww,
@drewww@social.coop avatar

Here's the hunter emerging behind you and chasing you. You really depend on abilities that move more than one space per turn now, otherwise the hunter will catch up quickly and end your mission.

Animation of a black, white, and red roguelike. The character emerges from the "elevator" and waits. Shortly, an "H" character appears and chases the player around. The player evades using some jumping abilities.

drewww,
@drewww@social.coop avatar

Spent some time trying to add "juice" -- animations and nice UI touches that make the game feel cool.

I had this idea of animating overlays but my core engine didn't support it nicely. I built my own transparent canvas overlay, but performance isn't particularly good so it doesn't look smooth. It was a bit of saga.

I'll ship the final version with this in, since I think it's better than nothing. But it's not good. Some canvas optimization would surely help if I have time.

jake4480, to indiegames
@jake4480@c.im avatar

Cool to see writeup on a game that's already on my wishlist, Cobalt Core. It's on Switch, I may grab it there. I can't speak to if I like it yet, but I'm guessing I probably will. The genre is "sci-fi roguelike deckbuilder", which sounds odd but check the screenshots at these links- I kinda get it. I think it looks great - I dig weird strategy stuff like this sometimes. Now I'm even more psyched to play it.

https://arstechnica.com/gaming/2024/01/cobalt-core-is-your-next-just-one-more-roguelike-deck-building-habit

Steam link: https://store.steampowered.com/app/2179850/Cobalt_Core

Switch link: https://www.nintendo.com/en-ca/store/products/cobalt-core-switch

drewww,
@drewww@social.coop avatar

@jake4480 its great! Lots of novel ideas and play patterns. Not quite enough cards to have slay the spire like depth but still good.

drewww,
@drewww@social.coop avatar

@jake4480 oh! yes definitely! The “original” though most of the mechanics come from Dream Quest.

drewww,
@drewww@social.coop avatar

@jake4480 Some others in this bucket that I've played and enjoyed: Monster Train, Roguebook.

drewww,
@drewww@social.coop avatar

@jake4480 some! Only one I’ve gone deep enough to get a win is Cogmind.

drewww,
@drewww@social.coop avatar

@jake4480 I have not! I tend to flit between games too much to really unpack these bigger designs. I also don't loveeee the melee combat of a lot of traditional roguelikes. Cogmind's dynamism there really captures me. Blowing up terrain, running away, outranging people, AOE, etc. I'll put those on my list to try though!

drewww,
@drewww@social.coop avatar

@jake4480 I…. Didn’t love loop hero. It’s a very clever design idea but it got repetitive. Too much “grind the meta progression” for me. I’m ok with light meta progression like card/item/weapon unlocks. But you can’t reason your way through loop hero. Lots of people like it though!

drewww, to random
@drewww@social.coop avatar

A reflection from writing performance reviews.

There are ~two approaches to earning trust. (1) Demonstrate similarity with someone. People like and trust people who are like them. (2) Be consistently useful to them.

The first approach is how most managers operate; they want to be perceived as "like" their team members and strive to be technically "legitimate." However this approach doesn't scale as you manage different kinds of people; you can't be everything to everyone in an authentic way.

drewww,
@drewww@social.coop avatar

The second approach DOES scale, but is harder. It requires you to be able to create change in the organization around you and deliver meaningful, scarce results. Money, scope, resources, or status. If you can deliver those, people with all backgrounds will come to trust you over time because you are useful and deliver on your commitments to them in ways they value.

drewww,
@drewww@social.coop avatar

@brainwane Oh that's a good extension! Safety can certainly contribute a lot to trust. Maybe even a precursor. Can you trust someone you don't feel safe around? Unlikely!

It's related perhaps to similarity. It's easier to feel safe with someone you perceive as "like" you because you expect they will react to circumstances similarly to you. And it's harder to develop safety without more common ground.

drewww,
@drewww@social.coop avatar

@brainwane I mean the latter -- do you have technical competence to manage them? Which is a mix of "can you represent my work accurately and make technical decisions in a way that makes sense to me?" and "do I have technical things to learn from you?"

A manager who can do neither is in a tough spot. But at a certain "altitude" doesn't need to -- they have other technically legitimate leaders between themselves and ICs.

drewww, to random
@drewww@social.coop avatar

Lets talk about "The Banished Vault" from Lunar Division.

You're in charge of a giant cathedral spaceship called the Vault. You're separated from home, and accidentally triggered a galaxy-eating phenomenon called the Gloom. It is chasing you. Your only salvation is to create monuments to be remembered by after it gets you. And envelopes your monuments, too. They are no real defense against it.

Your job: manage your people and resources to build 4 monuments and survive transit between systems.

jeffjarvis, to random
@jeffjarvis@mastodon.social avatar

The AI bubble....
The typical OpenAI engineer makes $925,000 https://www.bizjournals.com/sanfrancisco/inno/stories/news/2023/06/26/open-ai-is-paying-3x-the-industry-median.html

drewww,
@drewww@social.coop avatar

@jeffjarvis The PPU grants vest over time, so if it’s a 2 year grant we’re talking 600k/yr. Or 450k over 3 years. Closer to “normal” SF eng rates.

Also valuing PPUs for a private company is impossible. Who knows what they’re really worth. More likely they’re worth zero than actual stock because it’s easy to not be profitable but the company to still have lots of value.

My headline would be more like “OpenAI paying engineers in risky financial instruments to compete with public co. employers.”

drewww, to random
@drewww@social.coop avatar

The "enshittification" argument from Doctorow is descriptively accurate, but it's stronger if you lay out the "why." It's capitalism all the way down. As cash gets tight, the least efficient businesses (sub-scale ad supported social networks) all are getting squeezed. So they are turning the screws on whichever side of their marketplaces can take it to eke out profitability. It's not a moral issue for CEOs, as tempting as that is.

drewww,
@drewww@social.coop avatar

There are no easy or right answers in this environment, and "doing right by the user/community" is possible when you have margin to spare. Erika Hall's argument here resonates with me strongly: https://medium.com/mule-design/a-three-part-plan-to-save-the-world-98653a20a12f

josephseering, to random

1/ I think it would be fair to say that, historically speaking, Reddit leadership has not had an especially sophisticated philosophy for platform governance.

Ordinarily, this would be a problem, but a happy accident of the relatively unsophisticated philosophy they did have was that it drove them to cede significant authority over large sections of the platform to people who had a fairly sound philosophy for community governance.

drewww,
@drewww@social.coop avatar

@andresmh @josephseering

The question IMO is not 10x, it’s can they survive at all. The risk is they are only barely SUSTAINABLE if they have free/cheap labor. Agree we might be tracking towards a utility model with much smaller engineering footprints. More like tumblr, less like Reddit. Teams of 200 not 2000 and platforms iterating way less towards growth.

Also shifts the motivation to contribute. When it seems like someone else is getting rich, relationships are strained.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • megavids
  • thenastyranch
  • rosin
  • GTA5RPClips
  • osvaldo12
  • love
  • Youngstown
  • slotface
  • khanakhh
  • everett
  • kavyap
  • mdbf
  • DreamBathrooms
  • ngwrru68w68
  • provamag3
  • magazineikmin
  • InstantRegret
  • normalnudes
  • tacticalgear
  • cubers
  • ethstaker
  • modclub
  • cisconetworking
  • Durango
  • anitta
  • Leos
  • tester
  • JUstTest
  • All magazines
    •