@uwunetes mine is a roadrunner v3 and it already kinda is a moped, except it lays down flat in my trunk
English
Conrad Kramer
939 posts
Conrad Kramer
@conradev
I work on chatgpt desktop @openai, previously co-founder @SkyBySoftware, shortcuts @Apple. convivial optimizer
West Hollywood ↔️ SF Katılım Haziran 2009
2.3K Takip Edilen11.3K Takipçiler
they're introducing a bill in california (SB-956) that makes it so class 2 & 3 ebikes (20mph throttle or 28mph pedal assist) will require registration
it will also make it so ebikes are limited to 750W PEAK, with anything higher being classified as a "motor-driven cycles" making it a requirement to have a motorcycle license to ride it
both of my ebikes have over 750w peak but are 750w continuous lmao so rip
English
Conrad Kramer retweetledi
Wow, I totally disagree with this statement. At the current state, AI actually amplifies the developer to developer difference. If you were a 10x developer, you had good ideas + architectural clarity, this is a brutal advantage when using AI. Steering is a fundamental part of today's AI development.
Rohan Paul@rohanpaul_ai
Chamath on how AI agents are making the "10x engineer" distinction disappear because the most efficient "code paths" are now obvious to everyone. Just as AI solved chess and removed the mystery of the best move, AI is doing the same for coding, making the process reductive and removing technical differentiation. "I'm going to say something controversial: I don't think developers anymore have good judgment. Developers get to the answer, or they don't get to the answer, and that's what agents have done. The 10x engineer used to have better judgment than the 1x engineer, but by making everybody a 10x engineer, you're taking judgment away. You're taking code paths that are now obvious and making them available to everybody. It's effectively like what happened in chess: an AI created a solver so everybody understood the most efficient path in every single spot to do the most EV-positive (expected value positive) thing. Coding is very similar in that way; you can reduce it and view it very reductively, so there is no differentiation in code." --- From @theallinpod YT channel (link in comment)
English
@negroprogrammer and the gap is somehow widening and not narrowing
English
Can We Interest You in a $400 AI Keychain That ‘Behaves Like a Real Pet’? gizmodo.com/can-we-interes…
English
@mitchellh stalwart is a fine OIDC provider, but I actually run it all behind authentik with fleet and headscale. All on NixOS!
English
@mitchellh I am very happy with forwardemail.net -> imapsync -> stalwart -> stalwart-webmail
English
I'm dying for pi-mono-style minimal library that handles the hard parts of email (auth, syncing with local state, etc.) and gives me an opinionated way to add agentic loops on top of that. I want to build my own agents and logic and guardrails, I don't trust vendors right now.
English
@dankuntz @tylerangert this is a good distillation except also AppKit
English
@tylerangert Screens that could be a static png? SwiftUI. Literally anything else? UIKit
English
the only real heuristic you need:
1. gestures? fast scrolling? novel / unique animations? anything that needs to be 120fps? UIKit
2. static screens? simple, discrete state changes? SwiftUI
Jacob Bartlett@jacobtechtavern
CPU and thermal performance in a SwiftUI list explains why we had so many hitches. Even at rest, no scrolling, the CPU screamed at 100% capacity to render every gif, and well past 100% (distributing work across CPU cores) when I scrolled. The “Very High” energy impact rapidly heated my device. The measured thermal trace crept up towards Serious. When I wasn’t paying attention, the app was even killed by the OS, presumably hitting a critical thermal spike. After giving a serious beating to my shiny new A19 chip, the same feed in UIKit produced comparatively nearly trivial CPU and energy usage. * At rest, UIKit dropped as low as 11% CPU utilisation, vs a consistent 100% for the SwiftUI version * Energy usage correspondingly held at High for UIKit, vs Very High for SwiftUI. I gave up waiting for the thermal profile to hit Fair after 3 minutes. Read my scientific performance comparison right here 🧪 blog.jacobstechtavern.com/p/swiftui-vs-u…
English
@turtlekiosk LLMs! Fundamental components like parsers are easy to ground and build verification loops for. LLMs can hill climb for hours in a good loop.
English
i really expected everyone to start porting everything to rust and it didnt happen
beginbot 🃏@beginbot
it's so awesome how all our software has gotten so much faster, more stable, packed with useful features ever since vibe-coding revolutionized programming
English
English
i’m done being a promiscuous macOS user. time to settle down and find my linux machine
any good travel laptop recs for ml/gpu stuff?
alexine 🏴☠️@alexinexxx
@ishtwts fuck mac os, marry linux, kill windows
English
@Jess_Riedel @Winterrose Seconding this. or Los Altos rod & gun
but neither are technically “in SF” and idk if smoking is allowed 😛
English
@Winterrose Richmond Rod and Gun Club is a great place, although they sadly now require you to come with a member. I think becoming an associate member is still easy, but presumably it won't work if you're just visiting.
English
are there any places to smoke cigars and shoot guns in SF with cool people?
realizing that sounds fun
English
@tylerangert I think this is still “reviewing code”
“The most important thing to cover in a review is the overall design of the CL. Do the interactions of various pieces of code in the CL make sense?” google.github.io/eng-practices/…
English
trigger warning: new "core skill", more important than just "reviewing code" now is being able to really articulate what you are confused about in the codebase. so often these models get caught up in "best practices" and mangled instructions that they'll just implement stuff without stepping back to simplify. having a good code smell detector will take you a long way.
English
Conrad Kramer retweetledi
the two diseases eating SF right now:
1) existentialism - too many directions, paralysis of choice. you can build anything, go anywhere, so you go nowhere.
2) nihilism - why bother? the models will do it anyway.
the only founders immune to both are the ones who started with a cultural/moral intention - that's the only cure i've seen work.
the "make something people want" era is over. culture leads - tech follows.
Noah Smith 🐇🇺🇸🇺🇦🇹🇼@Noahpinion
Yep. But this time is different. When ChatGPT came out, everyone in SF was incredibly excited and upbeat. This time almost everyone is either anxious or hopeless.
English
@hdevalence same
when zig’s IO type shapes up I might use that for some stuff
English
i’m very happy with tokio
neural oscillator of uncertain significance@mycoliza
erlang was an incredibly important, influential language in the history of concurrent systems…but present-day Erlang users are really unwilling to admit that the actor model is just an alternative formulation of “locks own data”
English
@AshtonAU I’m biased and it’s free for me but I do use codex to run my entire life now
English
@drewocarr @MangoSweet78 what are you all using pip/uv for? Main codebase or like one-off tasks?
English
English
@KrauseFx the M4 Air beats that computer in multi-core geekbench (i.e. compiles)
English
I don't understand if it's worth getting the MacBook M5 Pro/Max.
I still have my Apple M1 Max 10‑Core CPU, 32‑Core GPU, 16‑Core Neural Engine (whatever this means exactly) 32GB memory
It costs at least 3k to reach the same performance, and 4k to be slightly above?
English