niko

Currently building platform products that help people book their next trip

[Eric Jorgenson 📚 ☀️ @EricJorgenson]

Everyone can use @elonmusk's "Magic Wand Number" and "Idiot Index"

They're universal ideas, helpful in any industry.

[Vox @Voxyz_ai]

https://t.co/r8JmqEuhlN

[Brian Armstrong @brian_armstrong]

Our new @CoinbaseDev portal is looking awesome! It's the best place to build in crypto.

It gives access to all our powerful developer tools (wallets, payment APIs, x402, etc) in one dashboard. You can even launch a branded stablecoin, backed 100% by USDC.

[Dan Shipper 📧 @danshipper]

how i hit inbox 0 every day with Codex:

[ElevenLabs Developers @ElevenLabsDevs]

Call your Hermes Agent

[Solomon Neas @solomonneas]

@AlexFinn If your agent keeps forgetting shit, make it write handoffs, review them, and save only what matters.

Brigade does it for you.

https://t.co/qa1Kmw0r4T

[Aakash Gupta @aakashgupta]

An OpenAI PM showed me his entire Codex setup, and the distance between how he runs it and how most PMs poke at it is the whole game.

Abhi leads international growth at OpenAI, where India just became the second-largest market and the fastest-growing one. His day starts before he does. By the time he sits down, three automations have already run: a Slack triage that reads every priority channel and surfaces what he hasn't answered, a growth dashboard that pulls seven or eight Databricks and Tableau sources into one web app and rebuilds itself every morning, and a weekly update drafted from Slack, Notion, Drive, and dashboards for his leadership group. He reviews and sends. The assembly happened without him.

None of that is the impressive part.

The impressive part is that he stopped writing PRDs. He builds the prototype first, then writes a ten-question FAQ as a companion doc. He takes features to 70-80% himself, points Codex at the most similar thing his team has already shipped in the repo, and hands engineers a pull request instead of a spec. The deliverable was always the product. The document was just how teams used to authorize building it.

Then he told me he built a 1040 tax-filing app to do his own return. He ran it against his accountant as an A/B test. The accountant's number came back with a higher refund, which didn't add up, so he sent over what Codex produced. The accountant had forgotten an income source. A model with zero accounting training caught a mistake a professional missed.

Here is what separates the two kinds of PMs now.

One kind uploads files by hand, writes a vague prompt, gets the wrong table back, and decides the tool is overhyped. The other kind builds the harness: every data source pointed at an exact table, a graduated permission model where reads and drafts run on full autonomy while anything going to a human gets their eyes first, and a skills layer where the person most tired of a repetitive task is the one who writes the skill that kills it.

The model is the same model everyone has. The harness is the part you build. And the PMs who build it are about to look like they have twice the headcount.

[Shann³ @shannholmberg]

how I’m building an agent company inside my agency.

the structure looks like this:

Agency gBrain
→ Orchestrator Hermes Agent
→ Department verticals
→ Specialist agents
→ Scoped sub-agents

gBrain is the company brain.

It gets ingested with the data and experience we already have:

> transcripts
> chats
> previous campaigns
> client learnings
> strategy docs
> internal workflows
> examples of what good looks like

That brain is maintained by a human champion plus an orchestrator Hermes Agent.

Under the orchestrator, we have different department verticals inside the agency.

Each vertical has its own specialist agents.

Some of those specialist agents have even narrower scoped agents underneath them.

I’ve found that narrow scope improves output quality and reduces drift.

> a general “marketing agent” is too vague.

> a lifecycle email agent with access to the right campaigns, voice rules, approval gates, and examples can get very good.

> a technical SEO agent with its own tools, checklists, and source standards can get very good.

> a content research agent with narrow inputs and a clear definition of done can get very good.

The narrower the job, the easier it is to improve the agent.

I use different harnesses for this.

Mostly Hermes Agent, but also CLI harnesses like Codex and Claude Code depending on the job.

I’m still looking for a good bare-bones harness for model routers to run on.

To keep track, I maintain an org chart inside the company gBrain.

The org chart shows:

> top-level orchestrator
> department verticals
> specialist agents
> scoped sub-agents
> which brain each agent reads from
> which tools each agent is allowed to use
> where human approval is required

For clients, I do downstream pods.

Think of them as new agent companies that are isolated from the agency brain, but can still communicate with our agency agents when needed.

A client pod has its own:

> client gBrain
> client orchestrator
> client specialist agents
> client-specific workflows
> client-specific approvals
> client-specific memory

This is important.

You do not want client context bleeding across accounts.

You do not want one agent with every client’s data, every tool, and every permission.

Scope is what keeps the system useful.

The powerful part is that once you build one vertical agent well, you can fork it.

Not copy-paste blindly.

You still need to customize the context, examples, approvals, voice, tools, and workflows.

But you are not starting from zero.

You might have 75% of the agent already done.

That changes the agency model.

You no longer need a full traditional department for every function before you can deliver a well-rounded marketing service.

One or two strong marketing engineers can run an output surface that used to require a much larger team.

But this only works if the agents are actually good.

It takes iteration, taste, source material, QA, workflow design, and real marketing experience.

Bad agents do not become good because you connected more tools.

Vague agents just create vague output faster.

TLDR:
> turn the agency’s knowledge into a brain
> turn repeated work into scoped agents
> turn each client into an isolated pod
> let skilled operators run the system

[Emanuele @emanuelecodes]

[🔊 Sound on]

I spent the last 3 days building an engine (it's gonna be open source) that lets you create articles like this: audio synced to the page with highlights, a magic cursor mode (you have to try it), interactive code snippets, checkpoints, audio plays based on scroll and much more.

Code dropping soon, preview in the first comment.

Cheers.

[Csaba Kissi @csaba_kissi]

5 tools you should know about

https://t.co/DXXFsZ2uNj - Open-Source Frontier Voice AI
https://t.co/H8aIdypLnQ - shadcn/ui components with motion
https://t.co/QZr2VaB3U6 - One page CV builder
https://t.co/hU497sy8Zu - Search 778 niche job boards
https://t.co/VrUtH6ZNse - Open-source Lightway terminal

[Hiten Shah @hnshah]

I’ve built and rebuilt Skills 1,000+ times across companies I’m involved in.

I’ve also helped 50+ friends build them for internal workflows and public use.

The biggest lesson:

Your company already has Skills.

They are sitting in old docs, Slack threads, customer calls, review rituals, onboarding notes, and the heads of the people who know how the work really gets done.

The method is hidden in the work.

A Skill makes it reusable.

This is the first article from what I’ve learned building Skills in the wild.

[PlanB @100trillionUSD]

AI (gemma4:12b) understands bitcoin better than most humans, and certainly better than most traditional media😂

[Chamath Palihapitiya @chamath]

https://t.co/wgVgAuenfx

[Thariq @trq212]

https://t.co/R6exTuF7P8

[Naval @naval]

Software platforms are going to be rebuilt for agent-first.

[Alex Lieberman @businessbarista]

Completely agree.

The way you turn AI into a compounding asset is by codifying your best ways of working and then democratizing it across your company in a way that wasn't possible before.

One of the more common requests we get from companies earlier in their AI journey now is to stand up their internal skill library & fill that library with skills that we build through interviews/sessions with top performers.

[Matt Van Horn @mvanhorn]

TL;DR of my new article: every Agentic Engineering hack I know.

This used to be vibe coding. Around last Thanksgiving it got good enough to become something real.

📝 The moment you have an idea → /ce-plan a plan.md, with Compound Engineering by @kieranklaassen + @trevin. Fuzzy? /ce-brainstorm first
🙈 Make the plan, don't read it. Plans are for agents
🧠 Use /ce-plan for your deepest NON-code work too (strategy, specs, research)
🎙️ Get voice-pilled. Talk, don't type (@usemonologue or @WisprFlow)
🪟 Run 4-6 tabs in cmux (@manaflowai), one task each
⌨️ New tabs open straight into Claude or Codex
📱 Remote-control every session + give your agent its own email address on @agentmail
☠️ Dangerously skip permissions. YOLO. It's my computer
🔀 /ce-work --codex: route the build to @OpenAI Codex without leaving Claude
🔎 Run @slashlast30days before you /ce-plan
🥣 @meetgranola everything. Drop the RAW transcript in, don't summarize
✋ You're the taste; the agents are the hands
🎬 Build video in the CLI so an agent can write it (@HyperFrames_ )
📚 Point your agents at your notes + memory (@garrytan's GBrain, @supermemory)
✈️ Work from anywhere: mosh, tmux, @NousResearch's Hermes + @openclaw
📄 Share a plan with a human in Proof by @EveryInc
🛠️ Anything you do twice, write a skill for it. That's the compounding part
🌟 Contribute to the open source you love
🔋 Never-sleep laptop + a battery brick in the bag
🤖 Printing Press by @ppressdev: CLIs that run real life. Tesla, groceries, flights
⚠️ Watch for AI psychosis. Touch grass, talk to the people you love, build things people want (even if people is just you)

🚀 The YOLO TL;DR: paste this whole article into your agent and tell it to make a plan to set ALL of it up, one by one.

[Brian Armstrong @brian_armstrong]

Make sure your business is ready to accept AI agents as customers!

[@jason @Jason]

they are trying to kill cursor and lovable… and every startup and application — as I’ve warned

Infrastructure companies eventually try to win the platform game, then they learn and take out all their partners on the app layer

[Cole @colderoshay]

a little late on this but my god…
incredible read.