Sabitlenmiş Tweet
wyn
979 posts

wyn
@wyn_eth
simpler the better building agent native web infrastructure on ipfs and base with friends
L2 Katılım Ekim 2014
411 Takip Edilen3K Takipçiler


Someone trained a swarm model on 3 years of NBA data and let it loose on Polymarket
The result: $1.49M. He didn't build a better model. He built a better crowd.
Here's how the system actually works.
He takes MiroFish - an open-source engine that simulates thousands of AI agents and feeds it raw sports data.
The actual building blocks:
> Player stat vectors: points, rebounds, assists, eFG%, usage rate across 3 full seasons
> Team form tensors: last 10 games, home/away splits, pace-of-play, defensive rating
> Matchup history: head-to-head records, positional mismatches, referee tendencies
> Injury probability models: medical staff reports weighted by recovery timelines
> Line movement tracking: where the sharp money flows before tip-off
Check how this wallet works:
@0p0jogggg?via=sales" target="_blank" rel="nofollow noopener">polymarket.com/@0p0jogggg?via…
MiroFish generates 4,096 agents - analysts, bettors, insiders, oddsmakers - each with different reasoning. They argue, form clusters, shift opinions. Consensus emerges.
He pipes that consensus into a 12-layer transformer trained on the full history of his 16,695 predictions.
The model compares MiroFish output against live Polymarket odds. When the gap exceeds his Kelly criterion threshold, it enters.
Lakers at 40 cents. MiroFish said 62%. One position: $190,823.
The edge isn't data. Every sharp has data. The edge is simulating how 4,000 fake humans process it - before the real crowd does.

BuBBliK@k1rallik
English

@ryanrhughes @jasonfried @Tesla I actually love it because im willing and able to beat them at their game. But for the avg consumer it’s a shit show
English

I loathe buying cars from any dealer. I’ve tried so many ways to make “normal” car buying not take 3hrs yet somehow, some way, it always does.
I’ve had the car picked out, price negotiated, literally ready to show up and pick it up and it still takes 3hrs.
On the other hand, I’ve managed to buy a house without ever setting foot in it and having a notary come to my house to sign all of the docs in 15min. 🤯
English

The last car we bought was a @Tesla Model Y. Painless purchase process. No salespeople, no showroom, no upsells, no games, no haggling, no pressure. Just a personal choice on my own time, and a simple few-minute process handled entirely via a clear and straightforward app.
The next car we're buying is from another brand. And holy hell, it feels like I'm going back in time. Salespeople, back-and-forth charades, pricing games, "when can you come in?" before the deal is finalized tactics, etc. And I'm still doing it all via email so I don't have to deal with the showroom antics. I've modernized the process as much as I can from my side, and yet it's the same old same old.
They don't even feel like the same thing. In one case I'm buying a car with all the baggage that comes with buying a car. In the other case I'm buying a Tesla with none of the baggage of buying a car.
This experience could make me lament this other brand, but what it really does is make me appreciate and respect the lengths to which Tesla has fully reconfigured the car buying experience. It's become effortless, like buying any other product. As it should be. A car is just another product.
Bravo.
English

@mmatthias @morganlinton this is awesome. my bin of cables can be understood! i actually love the idea of just cataloging all the bs tech shit i have in bins and boxes and seeing what's actually worth saving
English

try knapsack but for the cables in your home, have it figure out what cable it is and how much it costs to currently replace, how much volume it occupies and optimize for value so you can have a bag with only your most valuable cables (cost to replace/volume of space cable takes up) have it let you keep "bags" of cables that are optimized
English

@bedemiralp @snwy_me lol this is hillarious. he asked chatgpt and it gave him bad info ;)
English

@blakeandersonw @10x_apps im actually starting to think it's almost better if you dont have a background in programing or design as you're not jaded by the past
English

After the sale of Cal AI, I have invested the majority of my earnings in @10x_apps
We're hiring engineers to build...
> AI employees
> AI consumer devices
Baseline $150k + 2% compensation, negotiable.
Apply below (in-person NYC) ⬇️
English

@morganlinton i was in the same situ. claude is a WAY better designer out of the box than codex but codex is my powerhouse. with the help of @iainmbean's component.gallery and a handful of sites i love, codex knocks out some impressive shit now
English

Okay, while I love GPT-5.3-Codex, and think it probably writes the best backend code out of any model out there now.
I do have to say, Claude Opus 4.6 has a better design eye.
Here's the frontend that GPT-5.3-Codex came up with for CodeGrader vs. what Claude Opus 4.6 created.
I mean, they're pretty darn different 👀
(First image: Codex, Second image: Claude)


English

@denisyurchak i mean it is but your exposing the issue with oc so it's both ;)
English

@wyn_eth A lot of ppl pointed out it is more a codex issue than OpenClaw but idk
English

@wholemars i mean right? how tf are ppl so damn gloomy. i mean besides all the fucked up shit going on at every moment too, but ;)
English

How I have my Mac Mini set up for @Openclaw (and why I’m using it over a VPS now):
Setup:
• Dummy HDMI plug ($4.39 on Amazon): tricks macOS into rendering a real display without a physical monitor (I mostly work on my laptop without external monitors)
• LaunchDaemon: runs at boot, no login needed
• Sleep disabled: never goes down
• Auto-restart enabled: recovers from power outages on its own
Why Mac over VPS:
• macOS ecosystem access: native apps, Shortcuts, anything Mac-only
• Persistent logins, extensions, GUI automation
Also:
• Screen sharing still works: I can remote in whenever I need to
I didn’t buy the Mac Mini solely to run OpenClaw (I was previously using a VPS), but I figured I’d make the switch for the reasons above.

English

I'm never using any app again. My new stack is just my assistant and a coding agent
Most things for work that I can automate I'll rely on Ava to do. Most personal use-cases are going to be done by Ava in the background.
I don't ever want to search for files, or download apps to do a certain things. I will only ask Ava to do it.
Last week, just for kicks, I asked her to order coffee for the whole office, from a thread in Slack. 10 mins later she delivered (pun intended)
Personal AI is going to change so much in the next 3 months, it's crazy how fast this will move. And I'm not gonna be sitting on the sidelines this time. Small announcement coming tomorrow ... :))



English

ENS domainers: let's see the Google Metrics from @grailsmarket for your best @ensdomains names!
drop them in the replies and we'll like and RT the best ones 👇👇

English

I built the first eSIM service for OpenClaw 🦀✈️
Now my AI assistant gets me eSIMs when I travel
I text it on Telegram or WhatsApp, the eSIM is delivered via chat
It consumes super little data and works even on shitty airport WiFi
What happened to me a lot:
I land in Cairo
Airport WiFi is slow af
Uber doesn't load
Roaming is €12/MB
It takes an hour and a lot of nerves to load an eSIM app
Now I literally text my OpenClaw:
“bro need 10GB in Egypt.”
- It finds the plan
- I press pay
- eSIM delivered in chat
I click a link and activate it
All you need is OpenClaw and an API key for eSIMPal



English

@wholemars i mean it certainly is artificial generalized intelligence, but the next question is more important. how generalized and how intelligent?
English

this is fucking cool. dude trained an ai model on his macbook by hacking into apple's m4 chip.
- he hacked into apple's inference accelerator (found in m4 chip) which turned it into a mini transformer.
- 80X more efficient than nvidia a100 gpus
- cost = electricity to charge macbook vs. renting gpus for $10Ks
- can be used to train, finetune and run models locally
my only feedback is its probably slow af but so cool regardless - you can imagine future chips (m6/7) can train much quicker.
English

@Jacob_Naviaux @shawngorham no question industrial is the right choice if your goal is to be as hands off as possible
English

More I hear about industrial, the more I get interested in it. Seems like the easiest to manage and scale. Although, I do like the hedge of affordable housing multifamily against AI displacement. If AI truly wipes out as many jobs as people are predicting in the near future, everyone needs somewhere affordable to live.
English

If I started Real Estate over again:
1. I would look at my primary house as a quality of life decision for my family vs an investment
(that means I would have stayed in the 2nd home we bought vs moving 3 more times)
2. After my family was settled - I would have bought industrial and retail strip vs single house and small multi
I would have focused on small contractor yards, odd shaped industrial lots, value add small user industrial, small run down neighborhood retail I could take down on my own.
I was not meant for single and multi family - it was just the easiest to finance
English








