Data Engineering & Topology. American.
Austin, TX
Joined June 2023
- Tweets16.3K
- Following2.6K
- Followers1.3K
- Likes50.9K
Clint J. retweeted
I’m pretty disgusted with the product landscape. we have machines spitting out code and yet 99% of the interfaces I come across are basically stock anthropic ui shit. please spend some time making your product sexy, people like sexy things. the looksmaxxers were right about this.
Clint J. retweeted
we used to get phone calls on the house telephone
there were rules
there was an edicit
if you called during dinner time and your friends parents answered you were fucked
mainly because your friends parents were sensible, but also why tf are you calling at dinner time?
now I get push notifications to the shitty pager thing I hate in my pocket, the thing that's plugged in on the end table next to my bed
this is not acceleration, this is retardation
sf really needs some help with their branding and mission docs
Clint J. retweeted
Introducing Pixel - a library for building GUIs in the terminal
- 120 FPS pixel-perfect graphics
- Embeddable inside existing TUIs
- Built in chromium webview component
pixelgui.com
Clint J. retweeted
Where will it flood? The National Water Center’s new high-resolution flood prediction maps answer that for nearly the entire US population. Now in ArcGIS Living Atlas. @nwsnwc
🔗: bit.ly/4rFTz6D
Clint J. retweeted
Speech can survive being flipped backward every twenty milliseconds.
The same speech in backward pieces of 200 then 100 then 50 then 20 milliseconds.
The original comes last.
When do the words become clear?
A listening demo from @OrukLabs.
Clint J. retweeted
ok 31k of you bookmarked this so i built it
paste any recipe, or just a link, and it becomes one of these
recipetables.com
Clint J. retweeted
Introducing Staging with ManimGX
Video models don't keep track of where things are. A shot can look right and still put the actor on the wrong side of the frame, flip an eyeline, or move the key light between cuts.
So I built a Skill that does the planning first: a top-down staging map of the scene.
Here's an example of how I use this skill to generate a scene with Seedance 2.5.
Clint J. retweeted
dario begged me in dms to not do agent swarms at home but whatever, this is nemotron 120B A12B running 64 concurrents at a bit over 100 tok/sec each for an aggregate of almost 7K tok/sec on a single GB300 (Asus DGX Station)
believe it or not, the agents did NOT magically just start hacking the planet tho. weird.
Clint J. retweeted
Some major performance improvements for PyNNDescent are coming soon, especially for large multicore jobs ... here's a teaser:
Clint J. retweeted
A Claude Coder ran Claude 945 times to measure what practices save on tokens vs. which do not.
Most useful:
- /clear
- /compact
- Move instructions from CLAUDE.md to a skill
Least useful:
- High effort (vs. Medium)
- Haiku 5.5 (vs. Sonnet 5.5)
- Unused rules
Clint J. retweeted
LLMs are either not capable or not affordable for visual reasoning tasks at scale.
The Elorian team - one of Jeff Dean’s best reports who led Gemini, an Apple LLM lead, xAI lead, and a Harvard professor - just launched an Astra-level visual reasoning model, Glimpse, for 17x cheaper.
Watch it break down the flying Tesla video physics here.
Extremely critical for video inputs in manufacturing, robotics, gaming, content moderation and more.
Clint J. retweeted
It was so bad luv
They also lied about their data retention policy, I told it to revoke access, it tells u it did (then it didn't) I am gonna have my other klank send them a support email + generate a negative review every 10 minutes all weekend for wasting special K's time
Clint J. retweeted
This is INSANE.
I ran 4 AI agents building 4 different 3D worlds in Blender simultaneously.
16 vCPUs. 32GB RAM. 4.5 minutes. In Parallel
Total cost? $0.006. Not even a CENT
We’re making compute absurdly cheap with Workpaper.
Giving away 16 vCPU + 32GB RAM machines for FREE.
DMs are open