Co-founder of @telepathinc. Former Mozilla, Blockboard, Storium, Delicious. Multiclass computing enthusiast. Your next computer will be made of magic and joy.
Edge of the Lost Coast
Joined April 2008
- Tweets27.1K
- Following994
- Followers2K
- Likes10.8K
Pinned Tweet
TELEVISION is now open source and available to everyone. think: ChatGPT Space but runs on your agent and any model. we’ve been cooking and testing with alpha users, excited to finally share it with you all. the next personal computer is coming. LFG. television.run/
Disassembling and carefully packing up the @telepathinc Alto for the trip to Barcelona. I'm demoing it at @mozilla's MozFest on Wednesday. If you're there, swing by and check out my computer from 20 minutes into the future.
100% this. there is a reason so much of computing moved to the GUI. those who forget DOS are doomed to repeat it.
Eugenia Kuyda, founder of @wabi, on why chatbots are the MS-DOS era of AI, and why a new interface is necessary to make AI's full potential obvious to everyone:
"It always struck me as strange that people were using products like ChatGPT, Gemini, or Claude, but mostly for very simple use cases: for search, or to help them write something, help with homework."
"No one was really mentioning any of these exciting capabilities that we just kept seeing the models get, and so we felt like there must be an interface problem."
"There should be something simple for everyone, me included, to actually discover these amazing use cases that all these models already have."
"We got obsessed with this metaphor that the current chatbots are really the Microsoft DOS era for AI interfaces. And that there will be some sort of a Windows/MacOS moment."
"There is so much traction with chatbots... In order to unlock something else, you need a more exciting interface."
@ekuyda
next rev of the @telepathinc Alto personal computer, the Mark II. new specs include dual DGX Sparks, a raptor lake i5 for the harness, and a push-to-talk CB-style mic for voice control. currently serving GLM 5.3 Flash, which runs Television indistinguishably from regular GLM 5.3.
a little bit of the future is unevenly distributed on my desk. it's the most fun I've had with computers in years. and this is just the beginning!
I'm bringing this rig to Barcelona next week to give a demo on the future of personal computing at @mozilla's MozFest. a bit of a logistical challenge but should be a blast.
so, I returned the Strix Halo. stellar machine. strong specs & speed, great value (esp given the Spark price hike), and it just works. but at the end of the day I personally need the added memory+speed of two Sparks. if your needs differ, I still highly endorse. nice job AMD. 🫡
🤖 Made with AI
amazing video from my cofounder @rupertmanfredi about "personal software" -- not just talking about it but *showing* it in action using Television! 😍📺
A brief demo & meditation on personal software I presented at a @getdiffer generative UI meetup last night
Stephen Hood retweeted
When roon is positive about something in AI alignment or monitorability, I take that as a very good sign.
Interactive, data-backed, open source artifacts on Television:
look I know Spark is the hotness but it's also the $. for $3K (!) less, AMD's Strix Halo competitor gets you the same 128GB and solid performance. it's quiet and comes with everything from the ROCm-verse preinstalled. I'm getting 70 tok/s with Qwen 3-Coder. so far i'm impressed.
setting aside CUDA vs ROCm and some differences in prefill speed, the main downside to the AMD is that you can't chain them as powerfully as you can with Sparks. but if your needs only require models that fit in 128GB, who cares?
also, again: *$3K* cheaper.
RAM prices being so out of whack that we're actually moving backwards from DDR5 to DDR4 was not on my bingo card and now I'm curious if there are historical examples of anything similar happening before. theverge.com/games/1009140/r…
how dare you, as a Columbus native, I... yeah, I concede the point.
Probably my favorite quote from my @MTSlive appearance yesterday. On the relative power of agent swarms vs more intelligent models.
those of us cobbling together our own multi-GPU rigs to experiment with on-device models and open source agents: we're the new Homebrew Computer Club, whether we realize it or not. we see the first glimmers of a new computer, and we're not waiting for someone else to build it.
