jamwise

Safari is the new IE

Consumer solar has been a game changer, so doing the same for wind would be a huge innovation. But the reasons everyone doesn't already have wind generators on their roof is it's a much harder problem to solve. Not least of which is wind speeds are significantly lower closer to the ground. So I expect their 1kw output has a very narrow "optimal conditions" window. This might work if you live in a desert? Though given enough time with climate change...

There are two sides to this with companies automating hiring with AI as well, so the whole interface between employers and job seekers seems to be getting an ever thicker layer of bots talking to bots. I've heard stories of interviewees clearly reading responses off their screen in video interviews even.

So although it seems like an inevitable arms race, having built engineering teams in the past this is all really unsettling. I've worked in, and hope to continue to work in, a remote-only job. So much of my process in hiring relied on relatively subtle signals. It seems like in person interviews will still be able to provide those signals for now, but I have a feeling personal recommendations are going to become increasingly critical for remote companies in their hiring efforts.

This is all giving me flashbacks of Rehoboam in Westworld.

Appreciate the feedback, thanks for checking it out.

To be clear, LLMs were not used to design this game, just to edit my post talking about creating the board generator. And I do play it every day, today was definitely a tougher one, but the variety day by day is by design, which is pretty common in word games.

Mini Mot | Jul 1, 2026 ⣤s⣤y⣀e⣀b⣀i⣀o⣀u⣀t⣀o⣀d⣀

Max Mot | Jul 1, 2026 ⣿d⣿b⣿i⣶r⣤i⣤o⣤e⣤l⣤c⣤t⣤s⣤o⣤e⣤

The key to this game is to use your general intuition about English phonetics to deduce where a letter can't go among the available spots on the board. Kind of like Sudoku logic applied to a crossword. So in theory you can solve quite a few words you've never seen before because certain letter sequences just don't work in English, (the premise behind XKCD for example).

Thanks for the feedback. The premise of the game is you solve each individual word in order, so greyed out cells are places that aren't the next letter for any of the words right now. As you place the letters those cells open up.

EDIT: I used an LLM as an editor, and clearly that stank. So I just touched up my original draft and updated the page.

Yeah that's fair. I might dust off my draft and re-edit myself, I foolishly thought this would lead to a better post, and I guess am not as attuned to the AI smell as others.

Seems like others feel the same. I guess getting an LLM to help with editing was a bad choice. Thanks for the feedback.

Well it certainly was a wall. That's how I kept describing it to my partner, I was hitting a wall.

Sorry about the turn off ¯\_(ツ)_/¯ I tried to put my best foot forward by reading about prose, engaging story telling, and did use an LLM to help me edit and reword parts of the post. Either way, I appreciate the feedback.

Yeah no criticism of the devs here, it's a great project. I meant to point at Apple. Though I'm a customer and like many of their products. I do hate their rigid clearly profit driven walled garden.

Information theory has been a really fascinating topic to get more acquainted with. Not really related to the crossword, but I highly recommend 3Blue1Brown's video "Compression is Intelligence", https://www.youtube.com/watch?v=l6DKRf-fAAM

Yeah I understand. But there are plenty of good headphones that work with Android and include most of the features. That's kind of the point of Librepods right? To bring AirPods up to par for non-Apple devices and make them a viable option for Android and/or non-Mac users. So "this working" means those features.

If I wasn't certain Apple will do their best to patch every avenue to this working in the future it might be motivation to buy AirPods.

Fair enough. I appreciate the feedback regardless. Thanks for sharing.

FWIW I was aiming for this to be a "connections" level of difficulty. My sample size is small but the past week win rate is about 65%, which is actually very close to connections. But I also don't want to cause frustration, since a game is supposed to be fun. So I want to think through this hint idea, maybe it can help lessen the load and also give a fun dynamic of deciding when to use your hint, so I'm thinking you just get one and have to decide when to use it?

Really great idea. And not going to ding it for kicking off with an LLM. But it might be worth thinking about how you could build a community of human contributors in a future stage. I personally don't feel comfortable with handing AI-generated stuff over to the kids in my life, and I imagine that would be a big barrier for your project as many others share this sentiment. That's a shame because encouraging kids to think deeply from a young age is something I would deeply support.

There are tons of knowledgeable teachers out there, and they are very often motivated to help the next generations thrive and grow beyond their day job. I'm sure you could find people to help with editorial work on this version to give it a human stamp of approval, and involve the many talented and inspiring educators and communicators that are out there to help grow this into a true value for the next generation.

Either way. Good luck!

Thank you! Glad you liked it.

My family plays a lot of word games, some are board-game style and of course we all love the NY-Times catalog of puzzles. The idea for this game came to me when I played bananagrams, and lost, after having binged way too much sudoku earlier in the day.

I became obsessed with the idea of making this work, and it was way harder than I expected. The journey took me into the depths of confusability graphs and zero error capacity in information theory. I might write a post about it some day.

The idea was fresh in my mind because I watched this yesterday. Great video, the illustrations and intuition-building of the compressability of information was so good! I'm so grateful for 3Blue1Brown.

Reminds me of when I tried to use the library of babel as a data compression tool. It led me down a fun rabbit hole and was my first introduction to information theory.

The conclusion being that you basically need the same amount of data to represent the address of your data as the data itself, so it's not really effective at compression, just a fun thought experiment.

The cool part of this in modern times is that LLMs are basically a form of lossy compression that actually achieves the gist of what these tools fail at. Although it is lossy, and requires a massive substrate. This is related to the idea of AI/LLMs being a form of language compression.

Speculative proposal

I guess at least they're honest about it? lol

This is wild. How did anyone approve this architecture. You should never give your LLM privileged access the current user doesn't have access to. Even if you're not logged in the LLM's tool calls should only be able to access the same flow you would, as in: be able to send a password reset email to your own email! This is like if you had a password reset page for your profile and had a email field you could fill in to have it sent to any email LOL.

I don't know if I'll ever use this but that "Infinite View" is a lot of fun, I just lost 20 minutes before snapping myself out of a trance. Some really cool pictures in there.

There wouldn't be any cloud. Splats are still local, but all the lighting and texture are pre-rendered. The problem is they're not interactive, so they'd be good for a lot of the environment but your main character and other things that need to be interactive would need to use a different approach.

I think it's inevitable it goes there. Right now the level of detail and quality of games is limited by the console/PC hardware you're playing on. But with the splats they can render the whole game's world in a massive server farm at Hollywood Movie quality. I imagine there might be some balance of splat and traditional rendering technology since not all objects will lend themselves well, but this might be truly transformative.

I only interact with them through their word and logic games. They finally coerced me into subscribing, but to their credit it was a pretty good deal. Now I'm worried.

Evan has done really great work. I haven't used Vue extensively (not my company's stack) but am a huge fan of Vite and it has helped our React pipeline a lot. I've also recently started playing around with CloudFlare pages and workers and it's already such a pain-free process to get basic apps up and running, I imagine this collab will make my life easier.

"Small enough to run locally with just 16GB of VRAM or unified memory"

With many laptops dropping back down to 8GB because of the memory shortage there's some interesting pressures building in the industry.

It's frightening how often this happens. And these days with the boatloads of cheap computer and phone peripherals being bought every minute there's just no realistic way for an authority to monitor and regulate all of it.

I bet it's not an insignificant amount of devices out there that had their firmwares written by a "random small developer" who is in fact some kind of supply chain hacker.