Michael Growney

Spawning Agents

[Michael Growney @michaelgrowney]

Agentic Economy transaction volumes are incoming

Grok Bot + X Money for agent payment rails is inevitable https://t.co/KObV4JHs3N

[Financelot @FinanceLancelot]

The worst part about being ahead of the curve is, everyone thinks you're retarded.

[Michael Growney @michaelgrowney]

Excited to get Grok Bot in my Tesla

[Michael Growney @michaelgrowney]

@mikepat711 @elonmusk 🎯

[Michael Growney @michaelgrowney]

Robinhood Chain is so much better than Solana

[Michael Growney @michaelgrowney]

Roughly 58% of SMBs in the US use generative AI

Most SMBs are AI rookies with basic consumer-tier tools used for one-off tasks

Stop over complicating it. Build simple AI products that enhance those existing SMB verticals

[Michael Growney @michaelgrowney]

Grok Bot group chat is a game changer for adversarial agents

[Michael Growney @michaelgrowney]

Gemini 3.8 Flash is the model release most teams should pay attention to.

Better reasoning/coding at Flash speed & cost is where agents get practical.

The win isn't sending every task to the smartest model. It's matching model to the work..

[SkalskiP @skalskip92]

Gemini 3.8 Flash might be my new favorite Gemini model

- #1 image reasoning
- #1 data extraction
- #2 object detection
- 30% faster than Gemini 3.7 Flash

↓ Gemini 3.8 Flash deep dive

[Michael Growney @michaelgrowney]

Late 2025: Hugging Face turned down a $500M NVIDIA investment at a $7B valuation. Didn't want an investor big enough to steer decisions.

Less than a year later: acquired by NVIDIA for $12,930,300,000.

Nobody pays 86x for a P&L. They paid it for 18M developers and 3M models.

[Benji Taylor @benjitaylor]

You need to be @bot maxxing

[Michael Growney @michaelgrowney]

Vibe coding is a perpetual state of optimizing your agents toolkits

I have a daily routine that assesses new skills drops and models against my current loops

This helps me lock in tools I actually benefit from vs. wasting time configuring AI slop

[Michael Growney @michaelgrowney]

I rely on my Grok Bots more and more everyday

[Michael Growney @michaelgrowney]

Nothing's more annoying than hitting your usage limit because every call went to a frontier model

Most of your Hermes queue is routine. Let your Grok 4.6 / Fable 5.1 tag in a model like Llama to handle the simple stuff

[Michael Growney @michaelgrowney]

@yieldingbear Pretty much every vibe coded ai app can be powered by 1 of the 20+ free models on here

[Michael Growney @michaelgrowney]

Token maxing is turning into token optimization. Builders are relying more on routers than separate endpoints for each model now

Stripe is locking in transaction volume from the first wave of AI builders

Will be cool to see how they evolve to capture agentic volume on-chain as it grows

[Michael Growney @michaelgrowney]

How many low reasoning tasks are you running with frontier model usage costs?

I use Yielding Bear to easily switch those simple tasks to free/low cost LLMs and save my frontier model usage for the heavy lifts

https://t.co/xsvaSuezsw

[Michael Growney @michaelgrowney]

The model is not the moat anymore. The gateway is.

One model doesn’t fit every job. Separate configurations for multiple model providers is clunky.

The AI cost / quality benefit for builders exists in competitive inferencing from one gateway.

[OpenRouter @OpenRouter]

OpenRouter is joining Stripe.

We started OpenRouter with a simple mission: intelligence should be multi-model. Today, we are the largest AI marketplace & gateway, processing 10T+ tokens daily on 400+ models.

Joining @Stripe gives us the opportunity to accelerate that mission.

[Michael Growney @michaelgrowney]

The Grok 4.6 Price / Intelligence ratio beats other frontier models easily

[Michael Growney @michaelgrowney]

Dump your project goals into Grok Bot and ask it what bots to make

It doesn’t just set up a fleet.. Grok 4.6 looks at the mess and picks the lanes

It spun up 3 I actually needed

A noticeable difference in delegating quality > quantity for jobs compared to Hermes & OpenClaw