Brad Weber

@TheLinkPanda Get me in please

Are you in? If no, let me know. If yes, just share to your $LINK marine fren who isn't in it.

@CigsMake @CoachFleck3 So the movie Project X

The peak of American house parties was probably 2007–2011. We had texting and Facebook to get everyone there, so plans actually happened. You didn’t need a landline or a 3-day notice to coordinate. But smartphones hadn’t fully taken over yet. Instagram wasn’t documenting every second. There was no TikTok/snapchat, no constant filming, no fear that of acting retarded could destroy your life

@Jacobsklug I’m interest and am building an AI agency:

YC just announced their looking for AI-Native agencies. The agency model is about to split into two completely different businesses: A) Agencies that sell labor B) Agencies that sell leverage Only one survives long term. AI-native agencies don’t scale by hiring more people. They scale by building systems that replace people. The playbook looks like this: → Find a workflow clients already overpay for → Build an AI tool that does it 10x faster → Use services to fund development → Turn repeated work into proprietary IP → Eventually sell the tool, not the time The real shift: Agencies used to be talent businesses. Now they’re becoming software companies with cash flow. Most people will miss this window because they’re still optimizing delivery instead of building leverage. That’s the opportunity. I'm launching a community of like-minded builders trying to build their own AI-native agency. I'm going to share everything I know having built my own 7-figure AI agency. Looking for motivated people ready to learn & build. Drop a comment, I'll personally reach out.

I built a 16-agent AI team that runs my consulting business. They handle client follow-ups, competitive research, pipeline management, proposals, and code reviews. Every agent has a name, a role, and a personality file that defines how it thinks. Here's exactly how the whole system works:

@NoahEpstein_ I tried running Kimi-k2.5 through Ollama locally which needed over 1T of storage which isn’t feasible for most clients, and is why I asked which model you were using

@Bpweber Was just first to mind for free local models.

“Just get over it.” No. People died alone. Families were locked out. And it happened because ordinary people just followed orders. Forgetting is how it happens again. Remembering is resistance.

@PennStateMafia With what players?

I'm fine being the first Penn State account to have said 10+ wins next year and let's ride in the College Football Playoff. Others said we'd lose all of our guys. Said we were a laughingstock. It's all coming together. Whose joining the Mafia? CFP in 2026 #FloodgatesClosed

@marchant_kenney @JMUFootball Why?

The current transfer portal is beyond ridiculous. @JMUFootball fans deserve better than this nonsense from the NCAA.

Last snapshot before the Chainlink Reward Season 1 closes with one hour left to go. $LINK Folks Finance $FOLKS continues to give stakers the highest return at current cube allocations, with a Value per Cube of $0.000404. Mind Network $MIND now gives stakers the second-highest return, with a Value per Cube of $0.000370. Space and Time $SXT currently has the third-highest return with a Value per Cube of $0.000301.

Another update on the Chainlink Rewards Season 1 application that uses AI to allocate $LINK Cubes across Season 1 projects to maximize token rewards: With less than a week left for cube allocations, $FOLKS Finance, $Suku, Space and Time $SXT, and $Brickken continue to give stakers the highest return at current cube allocations with a Value per Cube of $0.000655 (FOLKS), $0.000522 (Suku), $0.000454 (SXT), and $0.000402 (Brickken).

Update on the Chainlink Rewards Season 1 application that uses AI to help allocate $link Cubes across Season 1 projects to maximize token rewards: Looks like FOLKS and Suku give stakers the highest return at current cube allocations with a Value per Cube of $0.000759. Currently running the application locally, but will publish it soon.

Season 1 allocations are entirely off-chain until the Dec 9 snapshot, so there's no on-chain data to track. The only source is the rewards.chain.link UI. Manually tracking 9 projects and doing the expected value math was getting tedious. Stack: React + Tailwind frontend, Supabase backend for storing historical data, pulls prices from CoinGecko API. The AI recommendations use Gemini Flash. Anyway, figured other $link stakers might find it useful. Let me know if you have questions about how any of the calculations work.

Other features: Historical charts show how cube totals have changed over time (helps spot trends). Sensitivity analysis shows what happens to your expected returns if totals shift ±10%. After the snapshot, there's a claim monitor that simulates the linear unlock schedule and early unlock scenarios:

I built an optimizer for Chainlink Rewards Season 1 that uses AI to help you allocate $link Cubes across Season 1 projects to maximize your token rewards.