co-founder @excalidraw

Joined December 2009
David Luzar (dwelle) retweeted
Moved the rendering of Mermaid over to Pele from @kepano. Same... but different. • ~450KiB ↓ • Faster • Prettier (in most cases) Also supports easy server rendering which will allow for more goodies. Stay tuned. We've slightly improved diagrams in general (as well as added Swimlane charts) so check it out.
4
20
463
13,316
David Luzar (dwelle) retweeted
Yapfast 0.2 is out! Server settings, polls, forums, push to talk and an AppImage for Linux, in a Discord client that still sits at about 40 MB in a call :D (sound on, it's a banger :p) yapfast.com/blog/yapfast-0-2…
5
6
295
David Luzar (dwelle) retweeted
We believe voice is the most natural way to interact with intelligence, so we created Dessa, the most authentic full-duplex speech model ever built. Talking with it feels like talking with a person…full of emotion, nuance, pauses, and a natural cadence. block.xyz/dessa
4,188
7,138
4,741
69,113
13,507,938
David Luzar (dwelle) retweeted
Introducing Yapfast (beta) I wrote my own Discord client because the normal one was eating 800 MB of RAM doing nothing; it's the same Discord, you sign in with your account, but it sits at around 40 MB. Free and open source, give it a go at yapfast.com :)
1
4
3
6
2,781
Absolute obliteration of the Pelican bench.
Argon has this amazing natural understanding of the world. It can do so many things from first principles. Here's a pelican riding a bike, using fourier epicycles. Video made by Argon. Time-synced Lyria 3.5 music prompted by Argon.
2
93
David Luzar (dwelle) retweeted
Excalidraw CLI released! Log in to your Excalidraw+ account or use with local files. You can easily render Excalidraw diagrams to give better output feedback to your agents. Install today.
49
97
25
1,485
73,714
David Luzar (dwelle) retweeted
Having trouble selecting a shape that's beneath another? You can now opt/alt + click to cycle through the shapes under the cursor.
1
3
62
3,880
David Luzar (dwelle) retweeted
Ok, auto-incrementing list items/counters is in! We're a bit conservative for now so it applies for numbers up to 42, and single-item lists only. Letters up to h, then switching to roman numerals. Newly, alt-drag duplicating a list item auto-selects its text so you can change it easily.
6
6
92
4,214
David Luzar (dwelle) retweeted
It's 4am. I just finished a project I've wanted for awhile. This is Slopalytics. It's a much better view for Artificial Analysis data. It also has a bunch of usage data from T3 Code. I plan to add much more in the near future, including some new benchmarks :)
193
67
22
3,111
266,513
David Luzar (dwelle) retweeted
Improved text editing experience. - When text tool active we highlight what you're about to edit or bind to. - Creating unbound fixed-width text inside containers now works properly. - The text caret stays visible as you type - we avoid UI by scrolling the canvas, and wrap text so it's not too long. You can always unwrap later if needed.
2
3
61
3,553
David Luzar (dwelle) retweeted
Improved our library a bit. Hover to show item names. Filtering gets you bigger previews and displays the item names right there.
1
4
56
2,712
David Luzar (dwelle) retweeted
147
773
126
6,994
368,286
Opus retakes the crown from fable on general reasoning (SimpleBench). We're getting closer.
3
227
hands down
Astra is still the best "review" model. It's so insanely thorough.
1
197
David Luzar (dwelle) retweeted
GPT 6 Astra Vs. Claude Opus 5.5...
8
11
1
309
28,170
David Luzar (dwelle) retweeted
Claude Opus 5.5 has the best visual design of any model I have tested so far
Claude-Pop - I'm Upping My P(Doom)
301
759
550
7,360
2,939,366
David Luzar (dwelle) retweeted
You can now add transition animations to your Excalidraw+ slides! And we have a whole new presentations editor. 🔥 What transitions beside sliding and fade-in would you wanna see?
4
10
89
8,135
David Luzar (dwelle) retweeted
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI-based economic revolution
4,028
8,283
6,820
76,632
40,132,792