newb

@danshipper laid out two architectures for agents in an org: 1/ everyone has their own agent that mirrors their taste and skills 2/ one super-agent at the center of the org that does everything for everyone I think for non-native AI companies, the answer is mostly 1, with a sprinkle of specialized task agents layered on top. here's why: → it adapts to tools and workflows that already exist. no forced migration to new AI-native workflows → gives every employee a super-powered intern instead of threatening to replace them → human taste stays the driver of output. the agent inherits the instincts of a specific person rather than generating the generic middle → decisions still have a clean human owner. "whose agent did/approved this" has an answer. this is especially important in risk-averse orgs → specialized task agents can plug in when a rigid process needs to be followed exactly → distributes the skill of working with agents across the whole team, not just one agent-whisperer at the center the super-agent model assumes you can build something that's good for everyone. but that's extremely hard in siloed workspaces where context and workflows are scattered and fragmented. yes, it's harder for every employee to "teach" their agent what to do, but I think the ideal scenario is that the agent learns through observing and not prompting. specialized agents generate the generic middle. employee agents carry the taste and judgement of the people behind them.

@varunram x.com/newbworks/stat…

Individual launch videos when someone moves to (or these days even visits) SF is the same pattern as 30 years ago when a wannabe model moving to LA would do pre emptive photoshoots and distribute them to all the modeling agencies (just like launch videos are distributed to VCs on X) Same pattern, just 30 years later and a city a little more north this time

@coldhealing God ❌ Clod ✅

Ladies, your future husband should be in mild ai psychosis talking to God through a large language model

at this point anthropic’s probably doing our jobs for us already but we don’t know it yet cause it’s unreleased

most dudes should obsess over their craft. the thing they can make, build, improve, or understand so deeply that it becomes their contribution to the world. roughly everything else is downstream of this.

@brian_armstrong this is on the first page of "things intense ceos say" - and it's timeless.

People are capable of far more than they think, on far shorter timelines. Problems expand to fill the time you give them.

@signulll from now on, I am no longer a shitposter. I am a witposter. 🫡

most ppl misunderstand wit which is one of my favorite qualities in a human & it’s surprisingly very rare. lots associate it with humor but that’s not the case. wit is the ability to notice a relationship between two things that should not obviously belong together, then deliver it before the social moment decays. that’s why it feels alive. wit is intelligence under latency constraints. as if opus or a deep thinking model had near zero latency.

@thedankoe based.

The greatest trait you can acquire is to work with tremendous intensity on things that matter to you, and more importantly, be strangely unbothered when those things don't work out.

@hunvreus i've literally been chasing the high of 12k likes on a tweet reply, and that's just me with like 45 followers - can't imagine what it's like to even have 1k followers, let alone 100k+ 😅

How the heck are people dealing with having 100k+ followers? I only had a handful of posts go in the 10k to 100k views, and it's really distracting. Tons of butthurt idiots insulting you, tons of AI replies, but lost in there a few genuine people actually engaging in a good faith conversation. What do you even do once that's what happens every time you post? Do you just ignore most of it @yongfook @levelsio ?

@levie no more making “ice” scores for new ideas when you can very quickly spin up experiments - seeing them through is another challenge 😅

Sorry to anyone who thought AI would mean we’d work less (at least for now). AI makes it easy to explore more than you did before, and so you start doing far more as a result. I regularly have seemingly small things that end up quickly consuming 3 hours because the agent made it easy to get started, but you still have to do the rest of the work to complete the project. This is work that I wouldn’t previously have handed out to anyone else, it’s just stuff that never got done because it took too long to do fully manually. And, counterintuitively, for some of these tasks as AI gets good enough at doing them, it even becomes economically worth it to hire someone to do it on an ongoing basis with agents. But until you could try doing them at a low cost you would never have tried. This is why AI won’t automatically reduce work in the way we imagine because work isn’t static. Most companies have far more they can do than they have today, it was just hard to get started on it all because of the natural constraints of time and labor availability.

@signulll the companies being listed - did they “lose pmf” or were they run poorly, failed to evolve, or something along those lines but then again these are possibly causes of losing pmf I guess?

which product in the market today has effectively lost product market fit after once actually having it?

@T_Zahil hmm have you tried going viral?

I’m thinking of quitting X There’s almost no real engagement left, no real content People comment with AI, like without reading, use schedulers to generate their posts with AI to get read by AIs And tbh I get pretty bad results considering the time I invest. I spent years of my life to reach 22,000 followers, and yet I don’t often get more than 5,000 impressions per post Sometimes the website feels empty, meaningless and uninteresting, and kills my motivation Sometimes I stumble upon good content or have a nice conversation, but it’s happening less and less X is less than 5% of Uneed’s traffic now, so I don’t really need it It’s just hard to say goodbye to something I worked on for years The lesson is clear: we don’t own our audience. The only way to be « safe » from algos and stupid billionaires controlling your reach is by starting a blog or a newsletter I’m going to give it 2 more months, because I want to document the Uneed Residency here, as promised But after this, if nothing changes, I think I’m going to leave. I’ll keep building and sharing everything I can, on my newsletter and my blog. Go to my personal website (Thomas-Sanlis dot com) if you want to follow my next adventures!

@hunvreus there’s gonna be a bunch of people here who’ll comment saying you’re wrong and that you’re an idiot and then they’ll go on to prove your point

Talking to smarter folks than me, I'm convinced many of the AI folks in my timeline are full of shit. Nobody is "running 20 agents over night" and building stuff for actual users. Maybe some are building internal tools or disposable software. Maybe. But building software people like using? That doesn't get hacked on day one or blow up after the 3rd user? Nope. I don't even understand what that's supposed to look like. Do you work out a 57 pages document that perfectly describes what you want to build and then summon 14 agents and have them run wild for 6 hours? And what comes out on the other end isn't a broken pile of shit? Nope. Not buying it. PS: it may also be that I have an IQ of 82 and can't figure it out.

...and with that, i've once again been rate limited by claude

@Rothmus ...from first principles

@pitdesi looks in terrific shape for a house built in 1885

Greg Brockman's apartment is for sale, where they started OpenAI I do NOT think there is an opportunity to build an OpenAI-themed Airbnb btw :)

tbh that's kinda what I had been doing too - but when I started reading about it, turns out only followers see the new posts that i'm posting - so basically no one's seeing what i'm writing cause I had single digit followers that's when I started trying to reply to relevant content to try to get people to start following

@newbworks @yashhq_22 Thanks for this man! I dont actually know how this space works. I’ve just been posting my content and hoping I get engaged one day. But will try this replying on people’s post. I’ve just been using my 8, 9 views to motivate myself to keep going 😂

What helped you grow on X faster? - great content - being a reply guy

x.com/newbworks/stat… there's nothing I can do that will ever top this achievement

@UxGsol x.com/newbworks/stat…

This guy literally hacked Polymarket with a hair dryer 💀 Happened in Paris. On Polymarket, temperature bets were settled using a single sensor near Charles de Gaulle Airport. He figured out the exact location, showed up in person, and placed a bet on an “impossible” outcome 22°C when the market expected 18°C. Then he pulled out a hair dryer and heated the sensor. The artificial spike got recorded as the daily high → market settled → he cashed out. Did it twice. Walked away with ~$34K. While everyone’s arguing over indicators and alpha, this guy is out here doing IRL market manipulation.