Clayton Kim (@TheClayMethod) profile picture
Joined July 2008
Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)Photo by Clayton Kim (@TheClayMethod)
Incredible piece. I felt personally attacked. It’s perfect
New blog post - Chasing the Karpathy Peak (and Falling Off Slop Cliff): Confessions from Slopaholics Anonymous My story of becoming a slopaholic, who’s most at risk, and how to avoid it (or recover) thomasjiralerspong.substack.…
1
156
This is the new 2nd Amendment
The only thing that stops a bad guy with an abliterated LLM is a good guy with an abliterated LLM.
1
97
After spending all weekend fighting Claude
How it feels to ask the Chinese ai models for help on something Claude won’t agree to
281
Can confirm. A little wonky since @muse webview doesn’t have Bluetooth enabled
exiting whoop after 18 months. kept the strap, ditched the subscription. it now runs fully local on zhoop (my fork of noop). no fluff metrics, my data stays on my phone. ai killed vendor lock-in.
181
We all have to become Product Managers
I understand the appeal in trying to find the first plausible fortress in our retreat from writing code, but if you think it's "architecture", I have bad news for you. The models are also very good at that.
193
This is how those gpt-4o people felt
a session you compacted 157 times is like a seasoned cast iron pan
2
469
Clayton Kim retweeted
me and my main agent watching a subagent do all the work
76
886
58
14,918
284,860
Verifiable reward can mean so many things. Super cool work
You can just RL a coding model to paint with javascript btw
560
The sheer craft and information density of all of his posts are astounding
New post 📝 For this article, I asked myself: what happens when motion becomes an input to our shaders? To answer it, I spent this summer experimenting with effects like motion masking, blob tracking, optical flow, and motion blur blog.maximeheckel.com/posts/…
3
499
Holy shit he's right. Incidentally a great signal for software engineering skill too
Let's all agree that this is the correct and final eval for agi jackhopkins.github.io/factor…
1
583
Clayton Kim retweeted
If you're in a verifiable domain, pivot to an unverifiable one.
119
246
138
3,203
727,204
I never understood this. Lettuce on salads is basically just a seed oil delivery mechanism
lettuce fans when they hear about a crop that has basically zero nutritional value and an extremely high chance of poisoning them and making them shit themselves
2
1,096
The timing and variability of the rewards basically match video poker
vibe coding is addictive because it’s basically a slot machine: -every prompt has an unpredictable reward -near-misses keep you one prompt away -constant sensory stimulation from token streaming, tests running, UIs changing -cost is obscured from you -the jackpot can be massive
3
681
It also revealed that a large number of people at companies are truly useless
Been funny watching people slowly realize these last six months that other peoples jobs are actually harder than they imagined. AI had everyone in March saying they could do it all. Then a bunch of PM learned about productionization of software. Engineers learned it’s nontrivial to get customers to like and use a UI. Designers learned that both PM and Engineering jobs are even easier than they imagined, but the amount of type A lunatics you need to argue with to get anything done collaboratively is insane. And now everyone is back in their own lane again.
650
Evals are all you need
The reason enterprises struggle to go beyond basic chat bots is the talent gap to build harnesses and evals. 1. Evals: do you clearly understand your use cases and can you replicate that in the form of offline and online evals. Do the evals express your ambition and do they push the jagged frontier of the models? Do they help you pick the right models on the quality-cost-latency curve? 2. Harness: do you have a system that manages routing, multiagency orchestration, context management, tool calling, memory that is independent of the models? 3. Talent: do you have the talent to build this all out on the frontier? This is the scarcest piece.
1
1,529
Member of technical staff
The ADHD dream career is being a highly paid consultant who swoops in, solves a complex crisis with a chaotic stroke of genius, and immediately exits before having to fill out the timesheets.
388
I have never been more proud to be Korean
South Korea has an adult population (aged 15–64) of 35.7 million people. 1 in 30 (3.4%) adults got margin called.
1
312
Everyone’s talking about using Fable or Sol as a planner, then farming work out to cheaper models. But who wants to have to specify each model for each task, including the verifiers? I wrapped the pattern into a reusable skill: github.com/theclaymethod/min…
2
343