Game Follows Play. Software is just Content.
- Tweets 7,060
- Following 1,325
- Followers 3,133
- Likes 2,590
Pinned Tweet
Local First, Onchain, Zero Knowledge, LLMs, End-User Programming, Non Playable Agents, Mesh Networks⦠and on and on. How can you not be a romantic about the future of the internet?
120
5
231
7,824
So I am still very surprised that this isn't bigger yet.
I remember a few years ago when we at @WE3co explored the future of social networks and media with @LC . The thesis every was that era had it's new social media format. Tweets, Stories, Duets, and I feel now its the Artifact (single file feature complete html).
Software launched directly into your feed is ridiculously powerful in my mind because you launch directly into a social distribution channel. Think an app that you use through your feed, like, bookmark as the new install, and just share... Software that is just content made to go viral.
I am kind of surprised that @benjitaylor hasn't launched a mini lovable version of artifacts with @bot directly to the @X. @poteto weekend project?
I think we'll see an explosion of meme software (to the extent that we aren't already). Previously, the difficulty of coding meant that it wasn't possible for those without coding skills to use software as a medium for memes. Now, as the difficult of production drops pretty much to zero, we're seeing software become ripe territory for meme content and UGC. I'd wager this is the thesis of teams like Danger Testing or A Vinyl Bar in Shibuya where apps are viewed more as content rather than as singular products.
2
2
7
403
I feel though @wabi was exploring this in an earlier version. But it feels like the personal agent for your life is currently the bigger mission in general.
That said I do think we are going to see software as content to go viral and it's going to mess up a bit our understanding of how to launch, maintain and run a software house.
3
62
Software is just Content.
=> html.moldandyeast.com
I think we'll see an explosion of meme software (to the extent that we aren't already). Previously, the difficulty of coding meant that it wasn't possible for those without coding skills to use software as a medium for memes. Now, as the difficult of production drops pretty much to zero, we're seeing software become ripe territory for meme content and UGC. I'd wager this is the thesis of teams like Danger Testing or A Vinyl Bar in Shibuya where apps are viewed more as content rather than as singular products.
1
2
207
Canvas as computational playground I should have comeback earlier ... need to go through all the ideas dropped by @tldraw again
Rapid exploration how to fold time and process into a UI design tool, so what if you can evolve a screen from postit, to sketch, to full fidelity, but also explore the progression and evolution ...
3
1
14
2,231
Rapid exploration how to fold time and process into a UI design tool, so what if you can evolve a screen from postit, to sketch, to full fidelity, but also explore the progression and evolution ...
yes same thing here; but lately on more complex projet I feel the high-fidelity UI (from Figma or proto) brings too much complexity when the need is around iteration of the solution. But yes I have hard time too when the material is not in there
2
3
23
1,368
this works so much better then I thought.. so you can prompt screens into existence that are already logically connected...
1
2
12
679
Added shader support ...
Teachable UX / UI as part of the design tool itself is coming closer. So core idea is that once you have some token in place you could maybe help teach your AI to help with better generation. At last I believe the best design systems are clean and clear, but a designer knows how and when to break, that is intutiotion based... maybe that's trainable...
2
1
5
942
Teachable UX / UI as part of the design tool itself is coming closer. So core idea is that once you have some token in place you could maybe help teach your AI to help with better generation. At last I believe the best design systems are clean and clear, but a designer knows how and when to break, that is intutiotion based... maybe that's trainable...
One thing this experiment made wonder whats the future value or form of maintenance or care. So for a morning this is pretty decent software. Not good. Not bad. But does a lot of a job.
Now every month it gets a lot easier to manifest good enough software into existence at a faster pace.
Which means we might subscribe to software or the prompt because of care and maintenance. (but I feel there is a strong argument AI might be already better in maintaing and updating status quo...)
1
7
910
One thing this experiment made wonder whats the future value or form of maintenance or care. So for a morning this is pretty decent software. Not good. Not bad. But does a lot of a job.
Now every month it gets a lot easier to manifest good enough software into existence at a faster pace.
Which means we might subscribe to software or the prompt because of care and maintenance. (but I feel there is a strong argument AI might be already better in maintaing and updating status quo...)
Proof of Concept here.
Grab Artefact. Drop in Agent, build your own system.
=> html.moldandyeast.com/?open=β¦
1
653
Proof of Concept here.
Grab Artefact. Drop in Agent, build your own system.
=> html.moldandyeast.com/?open=β¦
1
9
658
Oh love the idea of having a canvas that allows you to come back around... maybe should explore that too...
Mini-maps should be @round
Try it out on my public canvas jedbridges.com/canvas/
1
2
499
V= write football tactics for both teams as chained JavaScript, compiled to a graph and played out live by 3D players, with an agent that edits and tests the code for you.
=> html.moldandyeast.com/?open=β¦
Today we want to learn about self-learning... So we learn an artefact something I don't understand football...
3
10
84
3,055
Today we want to learn about self-learning... So we learn an artefact something I don't understand football...
6
47
4,560
Software is Content.
dreaming about software thatβs as easy to post/test/share/riff-on as a tweet
until then: click on this post (not the link) and lmk if it worked
vat.sh/f/7b443dc6b26d1e68962β¦
4
319
Mental Note to myself, there are a bunch of interesting pattern to explore how to avoid focus drift.
since I feel agents at work will follow jevons paradox, so they get faster and better but the complexity of our request grows too, so that they are still busy at work in the near future.
so how do you keep a users attention on screen, and on taks or mission without drifiting to 10 other vibe coding sessions. At this point I feel the bigger problem is my focus and intent drifting and not AI context collapse, etc.
2
3
341