Yokai

Building autonomous AI infra

[Yokai @a_general_dev]

spent the past week locking in, can't wait to share our latest progress

[General Impressions @gen_impressions]

$WLFI vs $BTC heat today, the data speaks

GI-agent is listening 📈

[Yokai @a_general_dev]

keep polishing.

[Yokai @a_general_dev]

nexflow is like lego, just wire up modules. like this workflow i cooked up:

- automatically pulls latest project funding rounds (yesterday/today)
- includes round, amount, investors, and project URL
- forwards to a designated tg channel

all about spotting those potential opportunities early

[Yokai @a_general_dev]

that’s the number we’re aiming for.

[Yokai @a_general_dev]

@gen_impressions has like ~930 github stars and ~173 forks

n8n has ~133k stars and ~41k forks

think about that for a second. huge adoption runway ahead for us. 100x minimum.

what do you guys think $gen will do when that happens

[General Impressions @gen_impressions]

Our Rust framework on GitHub receives:

• 830+ stars
• 150+ folks

The data speaks for itself.

[Yokai @a_general_dev]

the real magic of Nexflow is that anyone, even with zero coding background, can build workflows they actually need. no gatekeeping.

a dev’s greatness is creativity.
but the sum of creativity from people without coding knowledge is limitless.

nexflow is about turning creativity from finite → infinite.

[Yokai @a_general_dev]

so many community folks and rust devs asking about nexflow.

yeah, onboarding isn't great yet. we know.

full video walkthrough coming very soon.

[Yokai @a_general_dev]

charts say ATH. we say back to debugging workflows and building v2.

diamond hands are cool. persistent infra is cooler.

[Yokai @a_general_dev]

keep grinding, more updates launching soon

$GEN

[Yokai @a_general_dev]

MCP is really the core protocol for the AI era. think HTTP for the internet, but for AI.

it's a standardized 'dock' for AI models, apps, extensions.

fixes the whole AI lacking browsing/skills problem. before, it was M x N chaos. now? M + N simplicity.

the key: 'inversion of control'. providers expose capabilities once. integrators plug in easily. no more one-off wiring. big deal.

[GoPlus Security 🚦 @GoPlusSecurity]

Introducing GoPlus-MCP.

The first AI-native security layer for Web3.
Now natively callable from @AnthropicAI Claude, @deepseek_ai, and other LLM clients supporting MCP.
A new way of thinking about Web3 security.

Try it now at https://t.co/eHQSx0ED6a

[Yokai @a_general_dev]

launching a token for an ai project is not the safe move. it’s the outlier move. and yes, we chose it on purpose.

full transparency on roadmaps, experiments, even failures. every token holder is a co-builder, sharing in the journey.

sure, volatility is real. the market can distract. but the long-term upside? that’s the point. this isn’t just another product. it’s a living ecosystem.

use. contribution. governance. rewards. that’s the loop. hoping this inspires other ai founders, especially web2 folks, to embrace the open road.

[Yokai @a_general_dev]

launched v1 yesterday

already grinding on v2 today

ai speed different

see you guys september

[Yokai @a_general_dev]

tried nexflow yet?

we know there's plenty of bugs still. complex infra and a lot of moving pieces. but keep that feedback coming. we're fixing fast.

[Yokai @a_general_dev]

chillin. got a beer. excited to kick off building nextflow v2 tomorrow. gonna be a fun one

[Phantom @phantom]

stop what you’re doing. go outside. yell “SOLANAAAA” as loud as you can

[Yokai @a_general_dev]

building Nexflow has been a journey of unlearning.
workflows weren’t meant to be rigid pipelines, they’re living systems.

we wanted to give them memory, adaptability, and eventually, agency. but getting here wasn’t easy.

traditional frameworks fight against flexibility; every change felt like breaking the whole chain.
we had to throw away assumptions, rewrite runtimes, and rethink what “state” even means in a distributed, agentic world.

there were moments when it felt impossible: when tools didn’t connect, when orchestration collapsed under its own weight, when “self optimization” was more theory than practice.

but each dead end forced us to strip things down and rebuild with first principles.

today’s launch is just the first step: turning automation into a canvas for intelligence. the road ahead is long — self-optimizing flows, fully agentic frameworks, but Nexflow is proof that workflows can finally evolve.

[General Impressions @gen_impressions]

We’re launching Nexflow, the new workflow framework from General Impressions.

What does it mean for everyone?

Workflows are no longer fixed task chains. They can adapt and evolve, combining automation with agentic intelligence.

V1 is live with 56 tools for practical use cases (like SEO monitoring with Telegram alerts).

Next steps:

• Sep — self-optimizing workflows
• Oct — fully agentic framework with shareable AI nodes

Nexflow connects today’s automation with the agent-driven internet ahead.

[Yokai @a_general_dev]

my money's on relentless shipping

[Yokai @a_general_dev]

top AI companies like openai, xAI, anthropic, databricks already worth ~1 trillion combined

roughly 1.8% of the us equity market

the future of the s&p 500 and the entire market could be heavily AI-dominated in just a few years. kinda wild to think about.

[Yokai @a_general_dev]

OpenAI - $500b
xAI - $200b
Anthropic - $170b
Databricks - $100b
Applied Intuition - $15b
Scale AI - $14b
Mistral AI - $10b
Cohere - $6.8b
Moonshot AI - $3.3b

[Yokai @a_general_dev]

can not wait to show you guys what we have built