kasey klimes
banner
kasey.bsky.social
kasey klimes
@kasey.bsky.social
no man is an island

building + learning to expand human agency

📍brooklyn

kaseyklimes.com
the upcoming betaworks camp theme is the most exciting one yet—agent systems

I’m grateful to be a part of the betaworks community. there is no greater center of gravity for oddball creative founders thinking at frontier of possibility
We've announced our Spring '26 Camp theme: Agent Systems.

We're seeking teams building autonomous systems of agents that deliver complete solutions to concrete problems.

Read the full thesis: beta.works/camp
Apply! beta.works/agentsystems
Priority deadline is 12/24 🎄
December 16, 2025 at 5:30 PM
many of today’s problems with coding agents were foreseen decades ago
November 30, 2025 at 8:36 PM
relistening to user study sessions > podcasts about startups
August 30, 2025 at 12:24 PM
the rumors are true, hmu
Kasey is hiring! In NYC!
if you're an engineer in the new york area and the ideas in this post interest you, let's get coffee.

i'm hiring.

"When to Design for Emergence"

t.co/Q7zrQX3vYN
July 29, 2025 at 2:22 PM
holy fuck
July 23, 2025 at 3:53 AM
if you're an engineer in the new york area and the ideas in this post interest you, let's get coffee.

i'm hiring.

"When to Design for Emergence"

t.co/Q7zrQX3vYN
July 1, 2025 at 8:48 PM
I’d be very curious what the post:reply ratio is here vs twitter
June 20, 2025 at 5:00 PM
Reposted by kasey klimes
Saying LLMs are worthless/pernicious because they sometimes make mistakes & is causing a general decline in reading skills (true) is like saying Google Maps is worthless/pernicous just because it sometimes gives you bad directions & has caused a general decline in human navigation skills (also true)
June 17, 2025 at 11:20 AM
startups!
how has AI code-gen changed your team’s approach to collaboration? how have you adapted your team workflows? what are the second order effects?
June 16, 2025 at 1:05 PM
must be something like that
June 16, 2025 at 2:44 AM
woah, bunch of new followers here all of a sudden, did I miss something?

hi 👋
June 16, 2025 at 2:18 AM
hey laurel! st louisan (and fellow design educator) here, sounds like you’re in nyc, happy to grab a coffee and give you the inside scoop if you have time before you bounce. congrats on the gig!
June 16, 2025 at 2:17 AM
!
June 16, 2025 at 2:16 AM
ngl building a company is the hardest thing I’ve ever done
June 12, 2025 at 12:29 AM
I believe strongly in permissionless innovation and also that I specifically should not be allowed to code
June 9, 2025 at 3:36 AM
possibly, I’ve never heard that!
June 8, 2025 at 7:36 PM
kinda messed up that steve jobs is known for the "bicycle for the mind" metaphor when it was bill atkinson who built hypercard and steve jobs who killed it
June 8, 2025 at 7:14 PM
I am building end-user programming for AI agents because limiting what gets built to the tiny group of people with technical knowledge limits us as a species.
June 8, 2025 at 12:46 AM
I love the way this book vaguely feels like it was written for grade school students at the same time that it’s a universal reference point for the most intelligent people I know
June 3, 2025 at 12:16 PM
the week in pictures

(new york, I love you)
May 29, 2025 at 7:49 PM
bsky wednesday?
May 28, 2025 at 5:02 PM
6. do not present your solution as blocks. most do not look at a pile of bricks and see a home. most do not want blocks. instead, present the fully composed solution —then let its adaptability become clear through use. design for emergence is not a paradigm people come for, it’s one they stay for.
May 27, 2025 at 1:06 PM
5. build your own solution to each problem from your new blocks. pay close attention to where it feels strange. if you feel the urge to cludge new features on to your blocks, you may have the wrong blocks. think about how to address the need with better block design, not patches.
May 27, 2025 at 1:06 PM
4. decompose each solution into recursive building blocks until the same building blocks can be conjoined to solve all fundamental problems. size matters. too big a block and you won’t have good fit with the problem, too small and it will no longer be ergonomic in the hand.
May 27, 2025 at 1:04 PM
3. discern the corresponding shape of a solution to each core problem. what is the platonic ideal of an end user experience for each? map every force tugging on the experience—note where two push against the same spot. fold each pair into one design question and shape until both sit calmly together.
May 27, 2025 at 1:03 PM