“Data analysis is the gateway to scientific reasoning” generalizes to non-scientific domains as well imo
Kenny Workman@kenbwork
English
Sai
42 posts
what's the bizdork obsession with having people absorb each other's roles? if i had 10 people who could do the same 10 things equally well i'd still have them specialize
Gokul Rajaram@gokulr
DESIGN: THE FIRST AI CASUALTY I'm increasingly sure that 2026 signals the end of product design as a full-fledged stand-alone function within companies. If so, it will be the first role / function to be eliminated by AI on a go-forward basis. Instead of hiring FT designers, startups are hiring / will hire design consultants to create a design system that the founder likes (this takes a few weeks max). Once the design system is finalized, PM/Eng feed it into their AI tool of choice to generate prototypes. The design system is refreshed annually by the same consultant. Larger companies will likely not backfill design roles and will do some targeted attrition to reduce the design department to 20% the size it is today. If you're a designer, I think you have two choices: 1. Become an entrepreneur: Start a design agency and become the go-to resource for design systems for startups and even larger companies. This can be a good recurring revenue business. 2. Become a builder: Add PM/Eng responsibilities to become a product builder. Would suggest you embrace this proactively vs waiting for the other shoe to drop. I'm really sorry about this - some of my best friends and the people I admire most and have learnt the most from are designers - but it seems inevitable.
English
@yoobinray What do you think they should do instead? Something more focused on how computers actually work like CompE or EE?
English
Anything to keep students in school for an increasingly useless degree
krish@IamIronLAN
Stanford is kinda crazy because as a CS undergrad this term you’re choosing between: - CS336: 0 to hero on frontier model training - CS224R taught by Chelsea Finn (founder of Pi) - CS231N taught by Fei Fei Li (Imagenet, WorldLabs CEO) - CS221M Mech Interp Intro taught with Goodfire And a host of personal podcasts delivered by $T CEOs.
English
everyday i wake up and watch you guys act out one of these parts
judah@joodalooped
thread of types of reactions from programmers to LLM progress 1. the scaling law believer it’s all over, just a matter of time. might as well enjoy these last few years apps and UI will be a solved problem in a couple a years, we’ll maybe stick around as testing nannies
English
seems obvious but:
things that are changing rapidly:
1. context windows
2. intelligence / ability to reason within context
3. performance on any given benchmark
4. cost per token
things that are not changing much:
1. humans
2. human behavior, preferences, affinities
3. tools, integrations, infrastructure
4. single core cpu performance
therefore,
ngmi:
1. "i found this method to cut 15% context"
2. "our method improves retrieval performance 10% by using hybrid search"
3. "our finetuned model is cheaper than opus at this benchmark"
4. "our harness does this better because we invented this multi agent system"
5. "we're building a memory system"
6. "context graphs"
7. "we trained an in house specialized rl model to improve task performance in X benchmark at Y% cost reduction"
wagmi:
1. product/ui
3. customer acquisition
4. integrations
5. fast linting, ci, skills, feedback for agents
6. background agent infra to parallelize more work
7. speed up your agent verification loops
8. training your users, connecting to their systems and working with their data, meeting them where they are
English
@adele_bloch @dejavucoder Indoor bouldering is a good place for this. There’s a natural slowness between trying the problem, resting, and people mingling around
English
how to strangermaxx:
- best place to initiate convo is where people generally linger. this can be in line, at coffee shops, after a yoga class, etc. there needs to be an element of slowness so you can easily strike up a convo
- wouldn't recommend starting with 'hi' - that can be offputting and there isn't much to do that with that. you'll want an easier intro line
- my favorites intro lines are almost as though you're letting them into your inner monologue like:
- "I can't decide between X or Y (on the menu)"
- "have any favorites that you order here? I want to try something new"
- *comment on a product they're holding or wearing*
- "wow that yoga class was so great today - I have to come again"
- after you say this, you can usually sense their appetite for connection. if they feel warm and friendly, you can keep it going - "have you been here before?"
- typically the conversation starts flowing from here!!
- sometimes people aren't in the mood to chat and that's no biggie - you still got a nice quick interaction of the day. sometimes the conversations will go on longer and it can lead to a really wonderful interaction!
- i'd encourage you to stay a *little* extra longer than you typically would. fight through the initial discomfort so that you can increase your appetite for this type of convo - repeat this whole process a few times to increase comfort with this!
you'll start to realize that people are EXCITED to talk to you!
strangers aren't as scary as they seem!
and you'll start living a life thats more open and fun!
Pearl Rose@hipearlrose
you gotta be strangermaxxing you gotta be chatting up random people as much as you can. Big Individual lobby doesnt want you to, but that is why you must
English
I swear every couple of months I come across a Borges reference lol. I’m gonna sit down and read a collection of his works someday.
Also, good post
gabby@GabriellaG439
New blog post: "A sufficiently detailed spec is code" I wrote this because I was tired of people claiming that the future of agentic coding is thoughtful specification work. As I show in the post, the reality devolves into slop pseudocode haskellforall.com/2026/03/a-suff…
English
English
@MillionInt @shyamalanadkat I believe that one day we could, but not clear if we ever should.
English
The recipe behind today’s frontier reasoning models is surprisingly similar to AlphaGo:
1) Imitate large amounts of human data
2) Scale inference compute to reason better (back then it was Monte Carlo Tree Search, today it's Chain of Thought)
3) Use RL to go beyond imitation
Demis Hassabis@demishassabis
Ten years ago, AlphaGo’s legendary match in Seoul heralded the start of the modern era in AI. Its famous ‘Move 37’ signaled to us that AI techniques were ready to tackle real-world problems in areas like science - and ideas inspired by these methods are critical to building AGI
English
@poeticdweller Probably the same reason lifting is not cardio
x.com/hjluks/status/…
Howard Luks MD@hjluks
English
cocaine makes your heart beat fast which is what cardio does. so why is it bad for you? Riddle me that.
English
@tszzl @QuasLacrimas When was the heyday of rationalism? I remember reading HPMoR when I was younger
English
@QuasLacrimas don’t really think I was raised in the cult of rationalism frankly, it’s heyday had gone before I was a grown man
English
A common theme you see historically in intellectual cults like Freudianism and Marxism is that even after all the provocative and unique theories are discredited and predictions disproven, people brought up in the cult are so insulated that they labor under the impression that textbook cliches like “industrialization is intensifying and spreading around the world” or “humans are not entirely rational” are the original insight of their guru, so they instinctively fall back on “Well, maybe he wasn’t right about everything, but you have to hand it to Marx/Freud…” (Narrator: no, you do not have to hand it to them.) When aggressively questioned about this, they fall back on the temporary stranglehold their cult has established around some prestigious center of power (like clinical psychology, or third-world dictatorships). As this stranglehold slips through their fingers, they are left with nothing but academic appointments in literature and film studies departments.
I don’t think anyone would disagree that Yudkowsky’s cult has very roughly a similar relationship to Silicon Valley as the one Scientology has to Hollywood. There’s no doubt that Scientologists networking and pulling strings has led to some amazing and profitable movies being made. That doesn’t necessarily mean that your health will improve if you clear out your body thetans, though.
roon@tszzl
the rationalists writ large were mostly right about most things btw. if you instinctively snicker about yudkowsky, scott, or whomever i take you to be a fish who’s unaware of the water
English
Sai retweeté
Memo: What If We're Right?
We recently wrote a private letter to partners & friends of a common failure mode: the inability to consistently reason through the daisy chain of downstream consequence when non-consensus, low-probability, events actually occur
pages: 1-3
English
@rieszspieces if you’re in college pay attention to your dining hall cooks, your career depends on it.
English
if you’re in college pay attention in your real analysis class. your career depends on it
Craig Weiss@craigzLiszt
if you're in college, get an internship at either openai, anthropic, xai, meta ai, or google deepmind. your career depends on it
English
Feels like everyone making their own agent stumbles across the same primitives and thinks they solved something
Let me save you some time (read this, it's funny and useful):
- You're going to make an agent
- You're going to run it on benchmarks
> It's going to suck
- You're going to make a tool to analyze traces
- You're going to say this helped you
> It wont work
- You're going to think about role based agents for solving a single task
- You're going to make a workflow for solving a benchmark
> Both will work. Neither are generalized
- You'll think you made it
> It will be nearly unusable by an end user
> Back to square one
- You're going to realize you're stuck with a for loop
- You're going to think about swarms
> In swarms single agent usability doesn't matter
- But wait you need a task manager
- But wait you need a merge queue
- But wait you need compression for long jobs
> Compression is a foot gun
- But wait now you need an agent to manage it all
- But wait now you need something that checks to make sure the manager is managing
- You're going to go back to single loop agents
- Well, subagents seem like the way to do all of this
> Bam! Plot twist: subagents are hard to do well
- You're going to think "Hmm well subagents isolate context" because said so
- You're going to start to look at other agent implementations
> How have they all solved compaction, multi-agent, task management, memory etc.?
- You're going to realize it's all just tradeoffs, but most of them have only one side people care about
- "Oh it's all just context engineering"
> Yep. But it has to be good and it has to be general.
> Back to the starting loop. Rinse and repeat.
Congrats.
Keep it simple. Keep it general.
English
I ~never use special coding prompts or setups. Any hardcoded hacks that work now will be outdated/actively harmful within a few months. I just explain my rationale and recommended preferences, maybe point it to official system prompts, and let it work.
Minh Nhat Nguyen@menhguin
how I convey intent to autonomous AI agents: 1. frequently write 1-2 paragraphs of rationale (why > how) 2. when logging fixes, write Problem, Solution, Rationale 3. always make agents log direct quotes w/ no rephrasing 4. all instructions are recommendations subject to change
English
This is my main dilemma. How much time and effort should I spend expanding an approach that works/brings value now vs trying to get something that can potentially better leverage model capabilities now + in the future but will take longer/be harder to control?
Chris Lovejoy, MD@ChrisLovejoy_
English