pulling out this deep cut for all my non larpers out there

truly a generational run

i trust the x algo to show our pokemon cards to the right people

trying omarchy today. plan is to convert my old gaming pc (has an RTX 2070 super lol) into my local ai inference box. if all goes well, i’ll be able to fully control my house using this machine. mostly excited to do computer vision stuff.

found this aquapolis togepi reverse holo at a card show today, super minty. the vendor only comped using tcgplayer & priced it at 35. my brother in christ, you CANNOT only comp using tcgplayer. ALWAYS cross verify with ebay last sold. bought and flipped instantly, gg

some heaters that my bestfriend and I have picked up over the past few months. we’ll be vending our first show soon! you can find us on IG @ hipjumpers 🔗

I got invited and went to an “AI for marketing” presentation and my god, we are so unbelievably early

I’m genuinely so happy I didn’t let myself fall victim to the openclaw hype. No one even mentions it on the TL anymore.

This is not a complicated new feature. The easiest way to think of this is that you can make a “team” of people you work with (marketing, sales, engineering, qa, etc). You can chat with each profile individually, and each profile can hand off work from one agent to another.

i think hermes agent has achieved a perfect combination of being the most consumer friendly AND the most capable ai tool today. it's my favorite thing to share with people. "just tell the agent what you want to do, and it'll do it." idk how else to explain it. it just works.

gyms that have cold ass AC on blast are a disgrace to society. let me sweat damnit!!

so i recently noticed my usage costs were getting really high, even for models that are generally cheaper than frontier models. these costs were running up on projects that are pretty low stakes, like real simple stuff. i'm on the $20/month hermes plan, so i try to keep costs down as much as possible. i don't think i've ever used a frontier model for any personal work. it's just so damn expensive, and i'm not dumping all my credits to get 3% of the way through a task before hermes tells me to refill. i want as much performance and intelligence as i can get for as little money as possible. so i built this "bang for your buck" model evaluator and selector. it lets me assign models to my hermes profiles, which are each tailored to a specific job. it uses a free API key from artificial analysis — their benchmarks score where each model performs best — and an algorithm ranks each model to find the highest intelligence for the cheapest price. you can see deepseek v4 flash 0731 and gpt 5.6 luna dominating my lineup. based on my tool, they're probably the closest you'll get to frontier intelligence for most work, for as cheap as you can get it today. all built with hermes desktop and their plugin SDK, super easy

found an insane deal on facebook marketplace lmfao

Thank God for America. Happy Independence Day! Here’s to the next 750 years of American Greatness 🇺🇸

at some point, you must decide whether you will create value, or if you’re waiting around for someone to create it for you it’s all a game of makers vs takers