Explorer

Newest first — browse tweet threads

Keep on to blur preview images; turn off to show them clearly

Decisions of what kind? Sell? No sell? I fucking HATE political vagueposting dumb edging

Decisions of what kind? Sell? No sell? I fucking HATE political vagueposting dumb edging

We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1

avatar for Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
If you’re a man in your 30s, acquire some priceless jewelry. Go into debt if you have to.

If you’re a man in your 30s, acquire some priceless jewelry. Go into debt if you have to.

Author (Husk, Crypto Confidential): https://t.co/L2COrV5QA3 Building: https://t.co/HMcbuBhTP3 Teaching: https://t.co/Dy0FsZHQaz

avatar for Nat Eliason
Nat Eliason
Writing is one of the very few things I refuse to delegate to AI I let AI write code for me because the generated code is objectively better than mine, and I can still architecture it + manage + edit it to my liking I use AI for pretty much every other aspect of my life, and it does a far better job than the alternative But every single time I take a draft of something I wrote and I ask AI for feedback, it tries to turn it into the most neutered, corporate, insubstantial piece of writing ever created It wants to smooth out the edges. It tries to make it boring and inoffensive. It gaslights me into thinking what I said is not appropriate It's as if me having a personality and actual opinions bothered the AI. "This may alienate readers that..." yeah fuck that. It is precisely my goal to alienate those readers. How else would I be writing something worth writing if I don't offend half the people reading it? The only thing worse than asking for feedback on a draft is asking AI to write something from scratch. Even if you give it guidelines or reference material to "mimic my voice, style and tone", it just creates bland, cheesy, inauthentic pieces of writing. Three theories: 1) Writing is thinking. I'm not saying that you need to think _in order to_ write, I'm saying that writing _is_ literally thinking. LLMs are good at faking thinking, but they fail at critical thinking and having actual opinions. Maybe humans are just better than AI at this, no matter how good current LLMs get 2) LLMs average everything out. Because of the way they've been trained, they can only generate something approximating an average of everything they've seen. The average code file is good, the average essay is bad. On average, pre-AI code was of MUCH better quality than essays (not everyone can write code; those who can tend to self-select for high IQ; most published code works, verifiably so -- this is the opposite case for most writing on the internet) 3) I just love writing. I raise an eyebrow whenever I hear programmers say "I won't let AI write code for me, I love writing code and my code is just much better" because I think that's just ego. But maybe I'm the same for writing. Some programmers love writing code, and consider it their craft, and want to keep things manual and artisanal. Maybe I'm just the same but for writing What triggered this rant: I've been writing the docs for RailsFast these last few days. You would expect technical documentation to be somewhat boring and sterile, but I love writing, so I tried to make it better than that. I kept asking AI for feedback, and it kept trying to make it "enterprise-friendly", saying things like: "you could tighten it" and "this could backfire if you try to move upmarket". I couldn't care less about moving upmarket bro, I just care about saying what I think needs to be said For now, every single thing I write is almost guaranteed to be manually typed out by me. I'll still ask AI for feedback (challenge me, fact-check me, proofread, etc.), and I may even let it suggest edits, but I don't think I will be letting AI write anything for me in the foreseeable future A corollary of all this may be that learning to write may ironically be a moat in the age of AI

Writing is one of the very few things I refuse to delegate to AI I let AI write code for me because the generated code is objectively better than mine, and I can still architecture it + manage + edit it to my liking I use AI for pretty much every other aspect of my life, and it does a far better job than the alternative But every single time I take a draft of something I wrote and I ask AI for feedback, it tries to turn it into the most neutered, corporate, insubstantial piece of writing ever created It wants to smooth out the edges. It tries to make it boring and inoffensive. It gaslights me into thinking what I said is not appropriate It's as if me having a personality and actual opinions bothered the AI. "This may alienate readers that..." yeah fuck that. It is precisely my goal to alienate those readers. How else would I be writing something worth writing if I don't offend half the people reading it? The only thing worse than asking for feedback on a draft is asking AI to write something from scratch. Even if you give it guidelines or reference material to "mimic my voice, style and tone", it just creates bland, cheesy, inauthentic pieces of writing. Three theories: 1) Writing is thinking. I'm not saying that you need to think _in order to_ write, I'm saying that writing _is_ literally thinking. LLMs are good at faking thinking, but they fail at critical thinking and having actual opinions. Maybe humans are just better than AI at this, no matter how good current LLMs get 2) LLMs average everything out. Because of the way they've been trained, they can only generate something approximating an average of everything they've seen. The average code file is good, the average essay is bad. On average, pre-AI code was of MUCH better quality than essays (not everyone can write code; those who can tend to self-select for high IQ; most published code works, verifiably so -- this is the opposite case for most writing on the internet) 3) I just love writing. I raise an eyebrow whenever I hear programmers say "I won't let AI write code for me, I love writing code and my code is just much better" because I think that's just ego. But maybe I'm the same for writing. Some programmers love writing code, and consider it their craft, and want to keep things manual and artisanal. Maybe I'm just the same but for writing What triggered this rant: I've been writing the docs for RailsFast these last few days. You would expect technical documentation to be somewhat boring and sterile, but I love writing, so I tried to make it better than that. I kept asking AI for feedback, and it kept trying to make it "enterprise-friendly", saying things like: "you could tighten it" and "this could backfire if you try to move upmarket". I couldn't care less about moving upmarket bro, I just care about saying what I think needs to be said For now, every single thing I write is almost guaranteed to be manually typed out by me. I'll still ask AI for feedback (challenge me, fact-check me, proofread, etc.), and I may even let it suggest edits, but I don't think I will be letting AI write anything for me in the foreseeable future A corollary of all this may be that learning to write may ironically be a moat in the age of AI

built and sold @PromptHero, now making https://t.co/5nkYFCnfUL (AI SaaS template), and open-source 💎 Ruby gems

avatar for Javi
Javi
TrustMRR traffic has been steady for the past 2 weeks (30K visitors/week). My biggest surprises are the 3-minute session time and the Sunday traffic peak (all my other sites usually peak in the middle of the workweek).

TrustMRR traffic has been steady for the past 2 weeks (30K visitors/week). My biggest surprises are the 3-minute session time and the Sunday traffic peak (all my other sites usually peak in the middle of the workweek).

Data is public btw: https://t.co/HI3HevS4p2

avatar for Marc Lou
Marc Lou
Check out my interview with @dfriedmanWFED on @FederalNewsNet about how @USOPM is training senior executives to run a more effective government to better serve Americans: https://t.co/34dRcw2oGZ

Check out my interview with @dfriedmanWFED on @FederalNewsNet about how @USOPM is training senior executives to run a more effective government to better serve Americans: https://t.co/34dRcw2oGZ

Sign up for the program here: https://t.co/T9WN6BZBiI

avatar for Scott Kupor
Scott Kupor
Wow! This post is going places 🤯 I'll work on figuring out how we can all do this. I have to find a way to filter out those who are not serious. 🧐 Meanwhile, you can follow my own $1m journey on my newsletter: https://t.co/b9bS78Eo9B

Wow! This post is going places 🤯 I'll work on figuring out how we can all do this. I have to find a way to filter out those who are not serious. 🧐 Meanwhile, you can follow my own $1m journey on my newsletter: https://t.co/b9bS78Eo9B

I build stuff. On my way to making $1M 💰 My projects 👇

avatar for Florin Pop 👨🏻‍💻
Florin Pop 👨🏻‍💻