Focused on #AI #VR #AR #MR #XR #SpatialComputing | he/him | SWE @Meta | Previously @RealityLabs @amazon @alexa99 | Side projects @RGBSchemes | Opinions are mine
Seattle, WA
Joined June 2013
- Tweets11.4K
- Following1.2K
- Followers1.5K
- Likes15.3K
Pinned Tweet
I've been using a chunk of my spare time to build out an agent runtime that is much more flexible than what exists today ime. Leviath is what I'm hoping will jumpstart the conversation around how adding structure to agents can improve their quality!
Gerald McAlister retweeted
Hey folks, like a lot of people this year I've been affected by layoffs, and I'm looking for my next opportunity! I'm a Lead Unity Developer looking for a similar position. I have a decade of experience, mostly XR but in small and large teams, medical, simulation, etc. DM me!
Gerald McAlister retweeted
Congrats to the team! Amazing to run a model of Muse Spark Open's caliber and scale this efficiently, right from your desk. Mac Studio keeps pushing the boundaries of on-device AI!
Gerald McAlister retweeted
Muse Spark Open -- weights releasing today -- runs great on new Mac Studios thanks to work with @johnternus and Apple.
Hot take, there’s one more moat: taste, but only when it comes to precision of execution.
Most folks who use AI to copy won’t be able to see those minute details quickly, and most AI can’t copy it without copying the exact binary details. It creates a challenge (for the foreseeable future at least) where it’ll basically change the app away from what one groups consumers like.
The iPhone Duo’s screen-open/close animation is a good example of this recently. Others copied it at a high level quickly, but there’s a ton of other details missing. Thus, they’re not likely to lose any users to such an effort.
Longer term of course yes, those features will be copied as people and AI pick up on the gaps, but I maintain this will take more time than folks realize. It’s the same reason why building a game from scratch right now looks janky with only AI, while decomps look “fine” depending on the platforms compared.
Something I think most folks don’t realize is just how tiny of a niche this is right now. The key thing this misses is that most “normies” (I really hate that word btw) are looking for solutions to specific problems, not “personal software”.
Most folks on X have 0 idea what the actual average person’s problems are, I can guarantee you that.
I'd call this a bubble in a bubble
Tech people somehow still think normies are not spinning up personal software already
They're not doing it in Claude Code in a terminal like tech people for sure
But their AI chat app is already doing it for them without them even realizing they're building personal software, like generating PDF reports, full web app dashboards and even entire web apps
That will keep getting more advanced and soon ChatGPT will just spin up a Photoshop clone (but ofc only if they need it, preferring to just do their photo edits for them)
Everyone will mostly use personal ephemeral software inside AI chat apps I think
Some software moats will remain like social media (community), ecommerce (think like Shopify and Stripe), etc but they will just be used by AI to build the personal software with
So maybe that's the point here, become a tool agents use to solve their user's problems and charge agents for it (like many are saying)
AI is gonna act as an extreme accelerant for two types of builders:
- the “sloppers”
- the “perfectionists”
Neither of these are right or wrong, but it will bifurcate the software experience severely. The cost will come from what you pay and what your experience is like 👇
Perfectionists will find they can actually ship, and with more features to boot! The experience will be pristine, and copying will be hard. AI is great at the first 90%, but that last 10% is where perfectionists thrive.
The tradeoff is users will have to get used to being told no. Some users may spin off to their own versions.
People who value their time won’t deal with sloppy software, just like they don’t today. People who value extreme flexibility won’t tolerate being told no, just like they don’t today.
AI is an accelerant, and trying to say that all software becomes tailored to a user misses the point: users vary in their preferences, and not everyone wants min or max flexibility.
Gerald McAlister retweeted
luma.com/seattlexr this should be it
Game development previously had the issue of many gamers thinking multiplayer was just a toggle flip.
The new version of this is gonna be AI gamers who think you can just “ask AI” to bring it to their preferred platform, not understanding the tradeoffs that come into play.
Guess we're gonna see folks re-invent software engineering as they learn why unit tests exist in the first place 🙄
Something that I think is often missed, is that there's a large swath of people who's desires are to make their lives as difficult as possible. Personal assistants "can't solve their problems" because the entire goal is making things easier to get done.
It almost makes me wonder if there's a market for an assistant that actually makes things harder? 😂
Because you're (intentionally?) misunderstanding it. The point is to reduce all the layers of complexity of internet tasks down to a single chat interface message, not to add more.
It's a really simple idea: today, only rich people have PAs to handle all the boring logistical tasks of life. With AI agents, everyone will have this.
But for real: Many people intentionally look for obstacles because they get enjoyment out of complaining. These folks will come to be miserable as AI continues to spread, and trying to target these folks is a big waste of time. There's always going to be some small reason AI can't do whatever it is they want.
The real goal needs to be figuring out how to negate the impact of these folks.
I'm not sure why everyone is so excited for clones of games right now with AI. Clones have existed since the beginning of gaming, it's nothing special. Yes, AI makes it faster to make a clone, but it doesn't solve the hard parts of getting the little details just right.
I'll add too: if by "clone" you mean "pirate", then I would expect the pushback on this to be stark and not go the way you think it will. The results will hit the quality of games, not the distribution. Games cost a lot of money to make, and AI can either help reduce the costs for slop or increase the value at the same costs.
Gerald McAlister retweeted
This is a take espoused by people with a poor understanding of software architecture in my experience (I don’t know Richard so not saying anything specifically about him here, maybe he has more refined thoughts)
I find it’s usually an over reaction to “agents write bad unit tests”. And instead of digging in and understanding why and how to correct… they prefer to throw it out completely
The pendulum will swing the other way imho
