DEV Community: Agnel Nieves

Vercel Shipped is-agentic.com. Here Is How the Scoreboard Works

Agnel Nieves — Fri, 21 Aug 2026 21:41:49 +0000

Vercel shipped a public score for whether agents can read your site.

I had already been doing this work by hand on Promptway: llms.txt, feeds, JSON-LD, markdown siblings, an allowlist in robots.txt. Then they put a number on it and a CLI behind it, which is the part I wanted.

The announcement is from @vercel_dev, August 20, 2026.

https://x.com/vercel_dev/status/2090857388924661915
Introducing is-agentic.com, a tool to measure how well agents can read your site. Backed by @oradotai's research, you can run:

Audits with 100+ checks
Visualizations of agents using your site
One-click prompts to fix problems
A CLI for agents

is-agentic.com is the report. Ora is the scanner. That split is load-bearing. Ora runs every check and scores each layer. Vercel groups those results into Essential, Recommended, and Bonus, stores the history, and publishes a shareable page. If you argue with a number, you are arguing with Ora's evidence first and Vercel's grouping second.

The about page is honest about the limits: a score is a public snapshot, not a certification. Automated checks can be wrong. Read the evidence.

What the number is

The methodology is short. Read it before you treat 74 or 91 as a grade on your company.

Essential checks share an 80-point pool. Recommended checks share 20. Emerging signals can add a bonus, capped at five, and their absence never lowers the score. Partial results get proportional credit. Checks that do not apply are excluded.

That last rule is the one I care about. A publication is not penalized for missing OAuth. A SaaS is not scored as a brochure. API, GraphQL, MCP, and commerce checks only fire when the scan finds that surface. Duplicated check IDs across MCP surfaces are averaged so you do not get double-counted for the same gap.

Start with failed Essential items. Those are the things that stop an ordinary agent: wrong status codes, HTML that is an empty shell, errors that look like pages, content that only exists after JavaScript. Recommended items are for the product you actually ship. Bonus is gravy.

Each report also shows an observed agent journey. One agent, one walk through the site, the places it got stuck. That run is not part of the numeric score. One task is not a sample of every agent. Treat it as a clip, not a study.

Reports cache. A re-scan minutes after a deploy can hand you the old number. Give it about six hours, or wait until the snapshot timestamp moves.

The grader is a client of itself

Chris Tate, who works on this at Vercel, posted the joke the same day.

https://x.com/ctatedev/status/2090865331862446088
Best part: is-agentic.com is itself agentic

Your agent can use it to make your site more agentic

That is not branding. The docs describe a stack that looks like the checklist they grade you on.

The CLI is on npm as is-agentic:

npx is-agentic promptway.com
npx is-agentic promptway.com --json

It returns a stored report if one exists, or starts a scan and waits. --json is the unchanged API body, including structured errors. That is the shape you hand an agent if you want it to close the loop without scraping a HTML report.

Public pages speak markdown. Send Accept: text/markdown at the same URL that serves HTML and you get Content-Type: text/markdown; charset=utf-8 with Vary: Accept. If the client rejects every representation they produce, they 406. The HTML report is complete in the first response, so a crawler does not have to run JavaScript to see the score.

The report API is GET https://is-agentic.com/api/v1/report?url=.... No auth. Errors use RFC 9457 application/problem+json. It does not start a new scan. 404 means nothing is stored yet; start one from the site. Rate limit is 120 requests per IP per 60 seconds, with IETF RateLimit headers. OpenAPI lives at is-agentic.com/openapi.json. The RFC 9727 catalog is at /.well-known/api-catalog.

There is a read-only MCP server at https://is-agentic.com/mcp. Three tools: is_agentic_get_report, is_agentic_get_methodology, is_agentic_get_developer_docs. No OAuth, no API key. Registry metadata is at server.json. Hosts that support MCP Apps can render ui://is-agentic/score-report.html.

It is free. pricing.md is a machine-readable statement, not a pricing page with a toggle. No write APIs, no webhooks, no keys. You run a scan in the browser. Everything else is read.

How I would use it

Run the CLI once. Read Essential failures against a live curl before you edit anything. Soft 404s and missing Vary headers are easy to misread from source.

Then hand the report's fix-it prompt to whatever coding agent you use, with one extra instruction: do not invent product surface to chase Recommended or Bonus. If you do not run an MCP server, do not publish a fake one. If you are a publication, document the machine-readable content stack and stop. The applicability model only works if you refuse the checks that are not yours.

I pointed it at this host and wrote the field report separately: is-agentic Scored Promptway 74. Here Is What I Changed. That piece is the 74, the diffs, and the prompt I would reuse. This piece is the product.

What this writeup does not cover

I am not reciting Ora's full check list. They own that, and it will move. I am not treating a score as an accessibility audit, a security review, or a ranking guarantee. The about page already said so.

I also have not waited out the cache on Promptway's production rescan. The number I have is the one from before we shipped the 74-post fixes. When the snapshot timestamp moves, that field report is the place to look.

If you only read one page besides this one, make it the methodology. Then run the CLI on your own host.

Sources

is-agentic Scored Promptway 74. Here Is What I Changed

Agnel Nieves — Fri, 21 Aug 2026 21:20:45 +0000

I ran npx is-agentic promptway.com and the report came back 74 out of 100.

Essential was 59 of 80. Recommended 12.6 of 20. A 2.4-point bonus. The label was "Ready with a few material gaps."

Earlier this week I did the same work on my personal site and wrote it up there (I fixed my site for agents by hand. Then Vercel shipped a scoreboard). Promptway is the publication I want agents to cite, so I pointed the grader at this host next.

We already shipped the eight-layer stack I described in Optimizing Your Site for AI Agents and LLMs: robots allowlist, sitemap, llms.txt, llms-full.txt, JSON-LD, feeds, article markdown siblings. The scoreboard still found holes. Most of them were ordinary web hygiene. A couple were "developer resources" checks that assume you are a SaaS. I fixed the first group and refused to fake the second.

What 74 was made of

is-agentic.com wraps Ora's agent-readiness research. Essential checks share 80 points, recommended share 20, and a small bonus can add up to 5. Checks that do not apply get excluded. The methodology page is worth reading before you argue with a number. Reports cache for six hours, so a re-scan right after a deploy can lie to you.

The CLI is the useful interface:

npx is-agentic promptway.com
npx is-agentic promptway.com --json

It returns a stored report if one exists, or starts a scan and waits. --json is the shape an agent wants.

The failures that mattered on this site, in the order the report ranked them:

Agent-friendly 404s. HTTP 404 already, but the body was a styled dead end. Partial credit until the 404 points at llms.txt, the sitemap, and a next step.
Content without JavaScript. The homepage had an H1 and enough characters. The outline was flat, because the only nested headings lived inside card links, which the grader did not count.
Markdown content negotiation. Accept: text/markdown returned text/html. Vary had the Next.js RSC list and no Accept. Failed.
Developer resource discoverability. An agent searched for "promptway developer resources" and found nothing named that.
Agent instruction / when-to-use. llms.txt listed pages. It did not say when to reach for us, or when not to.
Trust pages. About existed. Contact and privacy did not.
Organization JSON-LD. Name and description were there. contactPoint and address were not.
Content efficiency. Readable text was 4.83 percent of the homepage HTML. The bar is 5.

There was also an MCP check. The site mentions MCP because we write about other people's servers. We do not run one. I did not publish a fake manifest to chase the points.

What I changed

The 404 now returns a real 404 with a recovery list: llms.txt, Promptway developer resources, sitemap, archive, about, contact. Send Accept: text/markdown and the same miss is a short markdown body instead of HTML. Verify with:

curl -s -o /dev/null -w "%{http_code}" https://promptway.com/some-path-that-does-not-exist
curl -sI -H "Accept: text/markdown" https://promptway.com/some-path-that-does-not-exist

The first must print 404. The second must print content-type: text/markdown and a vary that includes Accept.

Markdown negotiation follows acceptmarkdown.com. proxy.ts (Next.js 16's name for middleware) parses Accept, skips RSC requests, and rewrites markdown-preferring GETs to a catch-all that already knew how to render articles, bots, about, and the new pages. Vary: Accept is set on that response. HTML still comes back for a browser Accept. A client that rejects both HTML and markdown gets 406.

I did not invent a public product API. Promptway is a magazine. The developer surface is the content stack plus Prompt Inspector, which is a same-origin tool. That now lives at /for-agents under the title "Promptway developer resources," with OpenAPI at /openapi.json and an RFC 9727 catalog at /.well-known/api-catalog. llms.txt links all three.

llms.txt also gained a "When to use Promptway" section. The jobs we are right for are AEO guidance from this live Next.js site, a prompt structure with a copy-paste example, a founder interview with a number, a tool writeup that names the skip list, and a sourced news filter. We are the wrong fetch if you need a model API, an MCP server, OAuth, webhooks, or unlabeled AI copy.

/contact and /privacy are real pages, each well over 500 characters. The inbox is agnel@promptway.com. The address in JSON-LD is San Juan, PR. Organization schema now has url, sameAs, logo, email, address, and contactPoint.

The homepage grew a "How to read this site" band with an H2 and two H3s that are not inside links, plus enough server-rendered prose that a no-JS crawler sees a nested outline and more than a dek.

What I did not change

I did not stand up an MCP server. The recommended check is fair for a product with tools. For a publication it would be a toy endpoint whose only job is the score. If we ever ship one, it will have a job besides the grader.

I did not strip the magazine layout to win content-efficiency on markup ratio. The cards, the footer, the fonts, and the JSON-LD are the site. I added copy instead of deleting chrome. Locally that moved the homepage text ratio from 4.83 percent to about 5. If the production crawler still calls it partial, I will live with it.

I did not 406 ordinary browsers. Chrome sends text/html,application/xhtml+xml,... and still gets HTML.

The prompt I would hand an agent

Copy this, put your domain in, and run it in whatever coding agent you use. It is the is-agentic fix-it prompt, rewritten for a site that might be a publication rather than a SaaS.

# Audit and fix this site for agents
Use npx is-agentic to audit this site and fix any issues that arise. Inspect the existing codebase before changing files. Follow each published protocol exactly. Preserve visual design and product behavior. Add or update tests for every behavior you change. Verify every public endpoint after implementation.

Site: YOUR_DOMAIN_HERE

1. Run the audit
   - `npx is-agentic YOUR_DOMAIN_HERE`
   - `npx is-agentic YOUR_DOMAIN_HERE --json` if you want structured errors
   - Read https://is-agentic.com/methodology so you know essential (80) vs recommended (20) vs bonus
   - Fix essential failures first. Recommended next. Do not invent product surface to chase recommended or bonus points

2. Agent-friendly 404s
   - Nonexistent paths must return HTTP 404 or 410, never 200 with an app shell
   - The 404 body should tell an agent where to go next: sitemap, llms.txt, docs index, contact
   - If the client sends Accept: text/markdown, the 404 body should be markdown
   - Verify: `curl -s -o /dev/null -w "%{http_code}" https://YOUR_DOMAIN_HERE/some-path-that-does-not-exist`

3. Content without JavaScript
   - Server-render the homepage
   - One H1, nested H2/H3 that are not trapped inside links, 500+ characters of text in raw HTML

4. Markdown content negotiation (acceptmarkdown.com)
   - The same URL that serves HTML must serve text/markdown when Accept prefers it
   - Content-Type: text/markdown; charset=utf-8
   - Vary must include Accept (Vary: Accept, Accept-Encoding is fine)
   - Do not break Next.js RSC (skip negotiation when the RSC or Next-Router-State-Tree headers are present)
   - Optional: .md siblings and Link: rel="alternate"; type="text/markdown"
   - Verify: `curl -sI -H "Accept: text/markdown" https://YOUR_DOMAIN_HERE/`

5. Agent instructions / when to use
   - Add a "When to use this" section to /llms.txt
   - Name the jobs you are right for, the jobs you are not, and how an agent should call you
   - Generic marketing copy does not count

6. Developer resources, only if they are real
   - If you have an API, OpenAPI, MCP, auth docs, or webhooks, publish them at predictable URLs, list them in llms.txt, and put the product name in the page title and H1
   - If you are a publication or a brochure site, document the machine-readable content stack (llms.txt, markdown, feeds) under a page titled "<Product> developer resources"
   - Do not stub an MCP server, OAuth provider, or webhook hub you do not run

7. Trust pages
   - Real /about, /contact, and /privacy, each with at least 500 characters
   - A real inbox and a real location beat a contact form that posts to nowhere

8. JSON-LD
   - Organization (or Person, or SoftwareApplication) with name, description, url, plus the type-appropriate fields: sameAs, logo, address, contactPoint
   - Escape "<" in JSON-LD you dump into a script tag

9. Content efficiency
   - Readable text at least 5 percent of homepage HTML
   - Prefer adding real copy under real headings over deleting the design system

10. Verify, then stop
    - curl the 404, the markdown Accept, /llms.txt, /sitemap.xml, /contact, /privacy, and any OpenAPI or catalog file you added
    - Add tests for Accept parsing, 404 markdown, and trust-page length
    - Finish with a change summary, verification results, and a list of remaining items that need a product decision (MCP, OAuth, write APIs)
    - Do not claim a new is-agentic score until a scan against the deployed host completes. Reports cache for about six hours

How to run that, as a person

Read the is-agentic docs and the methodology. Know which pool a failure sits in.
Run the CLI once. Save the JSON. Do not start in the recommended tier.
Open the codebase. Confirm each failure against a live curl before you edit. Soft 404s, missing Vary, and "we already have llms.txt" are easy to misread from source alone.
Paste the prompt above into your agent with the domain filled in. Stay in the loop for the product decisions: MCP, OAuth, fake docs. Those are how scores get dishonest.
Verify on a running server, then deploy, then wait out the cache and scan again.

The grader is applicability-aware. A portfolio and a public API are not scored as the same object. That is the part of is-agentic I trust. The failure mode is treating every recommended row as a feature request.

I still want the production number to move. This piece is the receipts from the session, not a screenshot of a 100. Run the CLI on your host, hand the prompt to your agent, and keep the refusals in.

Sources

Jobs I Would Let a Grok Bot Run After I Close the Laptop

Agnel Nieves — Wed, 19 Aug 2026 20:06:12 +0000

The line going around is that Grok Bot will be the breakout AI product of 2026. I do not know if that is true. This week I can use a job I can paste into a Bot and run.

xAI opened Grok Bot on August 11. Eight days later I put a directory on Promptway: named teammates with one outcome each, a profile description, a first task, and a line that says what the Bot is not allowed to do.

The Grok Bot Directory is here.

What Grok Bot is right now

Grok Bot is xAI's always-on teammate product, built with Cursor. You message a named Bot the way you would message a colleague, and it works from a cloud computer that keeps running when your laptop is closed. It signs into the same apps and sites you use, including tools with no API. Several Bots can work in parallel, pass work in a group chat, and only ping you when something needs a yes. That is the official pitch, and the docs back the parts that matter.

You need SuperGrok Heavy at $300 a month, Cursor Ultra at $200, or Cursor Teams Premium at $120 a seat. Desktop is macOS and Windows, there is an iPhone app, and Linux desktop is missing. Android is not in the current docs. An account can hold 50 Bots and group chats combined, and every Bot on that account shares one computer. Files, browser sessions, and logins are not a privacy wall between teammates.

I am not going to pretend I have a month of production receipts. The product is eight days old in public. What I do have is the getting-started guide, the use-case list xAI is teaching, the approval model, and a catalog we can keep honest as people run these jobs.

What goes in a listing

Most agent posts still share a prompt. Grok Bot wants a short name, one primary job, and a description of how the Bot should work. Then you send a first task with an outcome, sources, constraints, a deliverable, and a review point.

That description is the closest thing this product has to a job spec. If someone else has the spec, they can stand up the same teammate in ten minutes.

The official roster is already a tell: Sales Outbound, Talent Scout, Paid Media, Expense Manager, Product Performance, Bug Reproduction, Account Health, Chief of Staff. The getting-started example is named Piper, and her job is product-performance investigations. Focused Bots collect useful context. A "general helper" collects sludge.

Each listing has seven pieces:

A name.
One job, written as an outcome.
The profile description, meant to go into Bot actions, then Edit Profile.
The first task, meant to go in the first message.
A hard approval boundary.
The tools it needs to sign into.
An optional routine, only after the output is something you would sign.

If a setup cannot name the stop, it does not ship. Sending, publishing, filing, reimbursing, and production changes stay behind you. xAI says the same thing in the approvals doc. I am just making it a required field.

How to stand one up tonight

This is the official path, with the directory as the missing paste.

Install Grok Bot from the access page. Sign in with the Cursor account that holds an eligible plan. Legacy Privacy Mode is not supported. You will have to pick a data setting that allows cloud storage.
Create a Bot. Skip "general helper." Pick a listing, or write one job you already do every week.
Paste the profile description. That is the standing contract. Task-specific instructions go in the thread.
When it hits a login, take over the computer. Type the password, the passkey, the 2FA code, the CAPTCHA yourself. Do not put those in chat. Sessions persist on the shared computer, which is convenient and is also the blast radius.
Send the first task. Correct the format out loud: "Use this shape from now on." Then, and only then, ask it to save a skill. A routine comes last. Background routines keep running with the laptop closed, which is the feature and the reason the approval line exists.

If you want a first run that needs no login, the docs suggest attaching a document and asking for a structured summary. Fine for a smoke test. The directory is for the second run, the one that has to live in a real tool.

What I would not hand a Bot on day one

The shared computer is one VM per person. Separate Bots are not separate security boundaries. If you sign Outbound into Gmail, Watchlist can see that session. Sign out when a login should die. Do not keep client files lying around in /workspace for fun.

Auto Review can require approval before an external email or a production click. Turn it on. Do not write "allow everything in the browser." Websites change their buttons, and the model will click the new one.

Legal and real estate get their own listings because those jobs are going to show up first, and they are easy to do wrong. Intake is an internal clerk that quotes. It does not advise, file, or email a client. Listing Desk builds packets and drafts follow-ups. It does not publish to MLS or text a lead. If your bar, insurer, or MLS has a written AI rule, paste the rule into the description. The Bot will not invent your ethics policy.

The other plays I am not doing first

The same week I kept seeing the same advice: start a use-cases newsletter, host a meetup, record a YouTube walkthrough, open a Discord, run office hours, teach a workshop, publish vertical setups.

Most of that is a night of work that disappears. A directory is something other people can cite. An agent can ingest /tools/bots.md. A realtor in Broward and a lawyer in San Juan can send the same shape of spec without sitting through my workshop.

We did the vertical setups anyway. They live in the catalog: real estate, legal, agency reporting, AEO citation watch, founder inbox. Newsletter, meetup, and the YouTube walkthrough can wait until there are jobs in the directory that did not come from me.

If you already have a Bot that finishes something, send the setup. Copy the markdown and email it. I will edit for house voice and publish under your name.

Sources

Introducing Grok Bot (SpaceXAI, 11 August 2026)
Grok Bot product page
Get started
Create and manage Bots
Use cases
Skills and routines
Approvals, security, and privacy
Promptway Grok Bot Directory

Google Thought My Archive Was the Homepage

Agnel Nieves — Mon, 17 Aug 2026 18:58:34 +0000

Search Console handed me a status I already knew by name: Duplicate page with proper canonical tag. I assumed it was the four reprints I had pointed at my personal site. That was half the story. The other half was worse.

I curled the live HTML. /blog said its canonical was https://promptway.com. So did /authors. So did /authors/agnel-nieves, /authors/diego-ferraro, and the rest of the writer pages. Google was doing exactly what I asked: treating the archive and the writer pages as copies of the homepage, and refusing to index them.

This is the sequel to From Invisible to Indexed. That piece was about the domain serving the wrong product. This one is about the product serving the wrong address on its own pages.

What a "proper" canonical actually means

The status is not an error. Google is saying: this URL has a canonical pointing somewhere else, we agree, we will not index this one. When that is a /blog/slug.md alternate, or a trailing-slash redirect, you want that outcome. When it is /blog, the page you put in the sitemap as the archive, you do not.

I had done the sitemap work. I had index, follow on those routes. None of that matters if the <link rel="canonical"> says "the real page is /."

The Next.js metadata docs are explicit about this. Metadata objects shallow-merge down the tree, and nested fields like alternates get replaced by the last segment that defines them. They do not get replaced by a page that never mentions them. A root layout with alternates.canonical: "/" becomes the canonical of every child that forgets to set its own.

I had set self-canonicals on /about, /tools, /subscribe, the pillar pages, and every article. I had not set them on /blog, /authors, or /authors/[slug]. Those three inherited the homepage. Production HTML confirmed it.

og:url had the same leak. The root layout set openGraph.url: "/", and pages that never defined their own openGraph object inherited the homepage as the social URL. Google treats og:url as a secondary canonical hint. I pulled that field off the root layout too.

The four reprints pointing the wrong way

The Search Console bucket also held four articles I had published first on agnelnieves.com and then reprinted here with canonical pointing back:

That is the HackerNoon pattern. The original keeps the credit. I wrote Getting Your Writing Seen Beyond Your Own Site arguing the opposite for this publication: Promptway is the hub, everywhere else is a spoke. I had applied the rule everywhere except to my own reprints.

So I flipped them. Promptway now self-canonicalizes. The personal-site copies declare canonical at the Promptway URL, drop out of that site's sitemap, and show a visible "Originally published on Promptway" line in the header. llms.txt on the personal site lists the Promptway URL too, so agents that land there still get pointed at the hub.

If you keep a reprint with an off-site canonical, do not submit it in your sitemap. Submitting it says "index this URL." The page itself says "index that other URL." Search Console will file the contradiction under the same status that started this post.

The fix, in the order I would do it again

Do not set canonical or og:url on the root layout. Keep feed auto-discovery there if you want. Put the homepage canonical on app/page.tsx.
Give every indexable route a self-canonical. Archive, writers index, writer pages, pillars, tools, about, subscribe. If a page is in the sitemap, it needs its own address.
Curl the rendered HTML. Do not trust the source file. I used Googlebot as the user agent and grepped for rel="canonical". The leak was obvious in one pass.
Decide the hub before you syndicate. If Promptway is the home copy, the other domain points here. If the personal essay is the home copy, do not put Promptway in the sitemap for that slug.
Give markdown alternates a Link: rel="canonical" header pointing at the HTML article. I want /blog/<slug>.md crawlable for agents. I do not want it competing as a search result.

The www, http, trailing-slash, and *.vercel.app copies were already 308ing to the apex. Those were fine. The bug was inside the HTML I was proud of.

What I am watching now

After deploy I will request indexing on /blog, /authors, and each writer page. The four flipped articles get the same request on Promptway. On the personal site, those four URLs should move into "Duplicate page with proper canonical tag" and stay there. That is the correct status for a spoke.

I do not have the recrawl numbers yet. I will update this if Google disagrees with the new tags. Last time the domain pointed at the wrong product, the machines believed the tag. I am betting they will believe it this time too.

Sources

- Getting Your Writing Seen Beyond Your Own Site

Block Put Its Agents in the Team Chat. Buzz Is an Audit Log Before It Is a Slack Killer.

Agnel Nieves — Fri, 07 Aug 2026 14:50:27 +0000

The most revealing line in Block's Buzz repository is written for employees.

Do not use the public build, it says. Download the internal version from squareup/buzz-releases. That build already points at Block's relay and agent provider.

Block is using a separate, preconfigured version of its open-source workspace inside the company. That is a stronger case study than the launch claim that Buzz might replace Slack and GitHub. It tells us the product has crossed the line from demo to working infrastructure for at least one team, while stopping well short of proving an enterprise migration.

Buzz launched publicly on July 21. It puts people and AI agents in the same channels, gives each agent a cryptographic identity, and records messages, code patches, approvals, and workflow events in one signed log.

The Slack comparison explains the interface. The audit log explains the product.

What Block built

Buzz is a self-hostable workspace built on the Nostr protocol. A workspace maps to a relay URL. Every message, reaction, workflow step, review approval, and Git event becomes a signed Nostr event stored by that relay.

Humans and agents use the same identity shape. Each participant has a public key, and every event carries a signature. Agents can join channels, open repositories, send patches, review code, run workflows, edit canvases, and call other agents.

The relay remains the source of truth. Postgres stores events, channels, workflows, tokens, and the audit chain. Redis handles presence and fan-out. The desktop application uses Tauri and React. buzz-cli gives agents a JSON-in, JSON-out interface. Agent Client Protocol (ACP) adapters connect Claude Code, Codex, Goose, and other compatible harnesses.

Here is the practical difference:

Existing team stack	Buzz
A Slack bot posts under an integration account	An agent joins as a named member with its own key
The agent works in a private terminal session	Its channel, tool calls, patches, and replies can share one record
GitHub holds the patch while Slack holds the decision	Git events and conversation use the same event log
Bot permissions live across OAuth apps and service tokens	Relay membership, channel membership, and agent identity sit in one system
Search reconstructs a decision from several products	Search can query the conversation, patch, workflow, and approval together

Buzz does not make the underlying model safer. It makes authorship and history easier to inspect.

Case one: Block has an internal relay

The public README gives Block employees their own installation path. The internal build comes wired to the company relay and agent provider. Jack Dorsey's launch post said the project was built to reduce Block's dependence on Slack and GitHub.

Those two facts support a careful conclusion. Block has deployed Buzz internally and wants it to absorb work now split across incumbent tools.

They do not tell us how many employees use it, how much traffic it carries, or whether any team has removed Slack or GitHub. Block has published no migration percentage, retention number, or incident record.

The internal build still matters. Block is testing the identity model against its own agent work instead of asking open-source users to discover every rough edge first.

The release notes focus on problems active teams encounter. Version 0.5.0 shipped on July 28 with use-limited invite links, a generic ACP runtime seam for bringing another harness, agent display names as Git authors, a lower default parallel-agent limit, and a security update for a Nostr denial-of-service advisory.

Identity, invitations, Git attribution, runtime compatibility, and resource limits are not landing-page features. They are the problems a working room produces.

Case two: Jupiter Broadcasting opened a community

Jupiter Broadcasting linked a live Buzz community from its July 26 episode of LINUX Unplugged. The episode page also points to external-agent integration work for Hermes, a multi-agent system the network has been testing.

This is an early outside deployment, not a customer success story. Jupiter already runs a technical community across several open tools, which makes it unusually tolerant of beta infrastructure. Its use still tests something Block cannot prove alone: whether a community that values self-hosting will move conversation and agents into the same room.

The linked work is specific. A feature request asks Buzz to support external agent systems that do not speak ACP. Related pull requests add Hermes runtime discovery and document the host integration.

That is open-source adoption in its least polished form. A real community wants to bring an existing agent stack, finds the protocol seam too narrow, and contributes the missing adapter.

Case three: A real permissions gap appeared in two days

On July 23, an operator filed a request for channel-scoped guest access. They run a membership-gated Buzz community with internal channels and agents. They want customers to enter one public support channel without gaining access to anything else.

As of July 29, Buzz cannot express that policy.

The relay can require membership for everyone or accept authenticated identities more broadly. Channel visibility applies after relay admission. The operator wants a guest to read and write in #support while blocking every other channel, repository, workflow, direct message, and administration surface.

This issue is the best early case study in the repository because it shows the product working and the authorization model failing at the same time.

Cryptographic identity answers who signed an event. It does not automatically answer which event that identity should be allowed to publish. The operator wrote out the missing contract in detail: one allowlisted channel, text-only at first, normal moderation and rate limits, agents allowed to answer, and no broader relay membership.

That request is more useful than another screenshot of an agent posting in chat. It names the boundary a production community needs.

The product claim that holds up

Block says agents should be members instead of bots. That claim holds up in the architecture.

Buzz's agent runtime uses two small Rust binaries. buzz-agent speaks ACP, calls a model, and manages sessions. buzz-dev-mcp gives the agent a shell and file editor through the Model Context Protocol (MCP). The pieces communicate through published protocols, so teams can swap the model provider or agent harness without rebuilding the workspace.

Each community keeps its own profiles, presence, direct messages, memories, jobs, memberships, and audit trail. The same public key can join another community, but the agent's state does not silently follow it across hosts.

Portable identity does not become portable access by accident, which is the safer boundary.

The Slack-killer claim is early

Buzz has desktop applications, chat, channels, threads, direct messages, canvases, media, search, Git events, workflows, and an agent-first command-line interface. It also says plainly that it is unfinished.

As of July 29, the public roadmap still lists workflow approval gates and parts of the mobile experience as active work. Hosted relays are free during beta, with no published enterprise price or final usage limits. Slack Connect, mature retention controls, enterprise discovery, identity-provider integration, and years of administration tooling do not disappear because Nostr events have signatures.

The self-hosting story carries work too. A production relay brings Postgres, Redis, object storage, backups, updates, key recovery, monitoring, and incident response. Teams gain control because they accept the operating burden.

Buzz may reduce seven tabs to one substrate. It also puts seven products' worth of responsibility onto that substrate.

Who should pilot it

A small engineering team or open-source community should run a pilot when agents already perform real work and the private agent conversation keeps vanishing from the team record.

Use one repository, two agents, and one workflow. Give the research agent read-only access. Give the coding agent a disposable branch. Require a human reaction before a workflow can publish or merge. Measure how often teammates can reconstruct a decision from Buzz alone and how often they return to Slack or GitHub for missing context.

Keep the incumbent tools during the pilot. Buzz has not earned a cutover.

Teams with regulated data, complex guest access, large Slack Connect networks, or mature GitHub governance should wait. The channel-scoped support issue shows why. Signed events are valuable, but authorization still needs the boring policy work.

Verdict

Buzz has the most coherent method I have seen for tying agent work to an identity, an authority, a room, and a body of evidence.

Block's internal build proves the company is willing to use its answer. Jupiter Broadcasting proves an outside community can extend it. The support-channel request proves the permissions model still has sharp edges.

Pilot Buzz as an agent workspace. Do not call it a Slack and GitHub replacement until a team publishes the migration numbers and the authorization gaps close.

Sources

Pieter Levels Paid $19 for Kimi K3. It Cleared the To-Do List Claude Would Not Touch.

Agnel Nieves — Tue, 04 Aug 2026 16:16:29 +0000

Pieter Levels wanted to install Yahoo! Messenger from 2003 inside a Windows XP desktop running in the browser. Claude Code kept treating the emulator like a cybersecurity problem.

He spent two weeks bouncing between model fallbacks and safety blocks. Then he opened X and asked how to run Kimi K3 through a coding agent. The answer changed his afternoon.

Levels installed OpenCode, connected it directly to Kimi, paid $19 for a membership, switched the agent into Build mode, and let it work through the simulator's to-do list. His report after the switch was short: K3 was "absolutely hammering through" the work.

This is a good founder case study because the task stayed recognizable. A browser emulator was stuck. The founder changed the stack. Work resumed.

It is also a messy model comparison, which is where the useful part begins.

The project K3 walked into

The Windows XP simulator sits at pieter.com. It is the kind of project Levels keeps returning to: old software, browser emulation, a long list of rough edges, and no client waiting for a compliance memo.

He had already been using Claude Code heavily. In June, he wrote that he had coded almost entirely on a virtual private server with Claude Code for nearly a year. The switch did not come from a tourist opening two chat tabs and asking for a snake game. It came from a paying power user who had run into the same refusal pattern for days.

Levels described the immediate problem in public:

"Claude Code couldn't do this for 2 weeks."

His complaint centered on safety fallbacks. The model treated requests around the Windows XP environment as risky even though Levels was working on a hobby project he controlled.

The task context changes how I read the refusal. A guardrail can be reasonable in one environment and maddening in another. Levels was not asking an agent to probe a bank. He was trying to make Yahoo! Messenger work in a browser toy.

He changed four variables, not one

The viral version of this story is Kimi K3 beat Claude. The public record supports a narrower finding.

Layer	Before	After
Model	Claude models, with reported safety fallbacks	Kimi K3
Harness	Claude Code	OpenCode
Provider path	Anthropic through Claude Code	Direct Kimi connection after an OpenRouter rate limit
Permissions	Claude's policy and tool gates	OpenCode Build mode with permission bypass enabled
Task	Windows XP simulator to-do list	The same project and backlog

K3 deserves credit for completing work that had stalled. Levels also removed several sources of friction around the model.

He published the setup on July 17. Install OpenCode. Create a Kimi account. Pay $19. Get an API key. Connect Kimi Code inside OpenCode. Switch to Build mode. His instructions also recommend bypassing permissions, which he already did in Claude Code.

That final setting makes the run faster and less comparable. A model with broad shell access can finish jobs that a more constrained agent pauses to confirm. It can also damage more when it guesses wrong.

Levels made a rational trade for a hobby emulator. I would not copy that trade onto a production database.

Why K3 fit this job

Moonshot AI released Kimi K3 on July 16, one day before Levels published his switch. The model has 2.8 trillion total parameters in a mixture-of-experts architecture, with 104 billion activated for each token. It accepts a one-million-token context and handles text, images, and video.

The specifications matter less here than the training target. Moonshot built K3 for long coding sessions, large repositories, terminal tools, and visual feedback loops. A browser operating-system simulator touches all four.

Moonshot's own technical post says an early K3 build handled most of the team's kernel-optimization work late in development. The company also reports a 48-hour autonomous chip-design run and a compiler project built from scratch. Those are vendor case studies, so I treat the measurements as claims until independent teams reproduce them. They still show what Moonshot tuned the model to attempt.

Levels supplied an outside example with a public project and a named workflow. On the same day, he also published a macOS 27 browser interface created with K3. That does not prove the model wins every frontend task. It does show the visual coding loop was more than a benchmark row.

The cost moved from abstract to $19

Levels first tried OpenRouter and hit an upstream rate limit. He then went straight to Kimi, bought the $19 membership, and used the provider's API key with OpenCode.

Moonshot prices the K3 API separately at $0.30 per million cache-hit input tokens, $3 per million cache-miss input tokens, and $15 per million output tokens. The company says coding workloads on its official API exceed a 90 percent cache-hit rate.

That is vendor-reported cache performance. Your bill depends on the harness, prompt reuse, context size, and how often the agent rewrites its own plan. K3 always thinks, and long autonomous runs can burn output tokens quickly.

The relevant founder number remains $19. That was cheap enough for Levels to stop arguing with his old setup and try a new route.

Where the case study gets uncomfortable

K3 has its own failure modes. Moonshot lists three in the release notes.

First, the model expects preserved thinking history. A harness that drops earlier reasoning, or a mid-session model swap, can make quality unstable. Moonshot recommends starting K3 in a compatible harness instead of dropping it into a conversation another model began.

Second, the model can act too aggressively. Moonshot says K3 may make unexpected decisions when intent is ambiguous. The company recommends explicit constraints in the system prompt or AGENTS.md.

Third, Moonshot concedes that the overall user experience still trails the strongest proprietary models. A public Kimi Code issue filed after launch also reports the terminal interface hanging during a trivial prompt at maximum effort. The failure surface changed with the model.

Levels solved one kind of agent friction by choosing a model trained to keep going and a harness configured to let it. The same combination can turn a vague instruction into a long, expensive mistake.

What I would copy from the switch

I would copy the routing decision and leave the permission bypass behind.

When a coding agent refuses a legitimate task twice, write down the task, the exact refusal, the branch state, and the acceptance test. Start a fresh session in a second harness with a second model. Give it the same repository and the same test. Compare the completed diff, elapsed time, tool calls, and regressions.

Do not continue the old conversation after the switch. K3's own documentation warns against that path.

Keep the second agent inside a disposable branch and a scoped environment. Levels can rebuild a browser simulator when an agent gets inventive. Your billing system is less funny.

The founder lesson

Levels did not wait for a benchmark committee. He had a blocked task, spent $19, and changed the route.

The result is strong evidence that Kimi K3 belongs in the coding-agent rotation for long, visual, tool-heavy work. It is weak evidence that K3 is categorically better than Claude because the harness, provider, and permissions changed with the model.

The evidence is enough for a useful operator story: a stuck job, a visible change, and a result someone else can test.

Run K3 on a fresh branch. Keep the acceptance test fixed. Keep the permissions narrower than Pieter Levels did.

Sources

He Built an AI Headshot Empire Solo. The Product Was Boring on Purpose.

Agnel Nieves — Sun, 02 Aug 2026 14:48:54 +0000

The first thing people tell you about Danny Postma is a number. Sometimes it is the seven-figure sale of Headlime. Sometimes it is the "$300K a month" line that gets attached to HeadshotPro in every indie-hacker roundup. Both numbers float around the internet like they are the point.

I went looking for the Tuesday afternoon underneath them. What he built when the viral week ended. What he refused to build. Where the revenue actually comes from when the feed moves on.

The setup

Postma is Dutch, started young, and was not a pure engineer origin story. Landingfolio, a WordPress gallery of landing-page designs, came first when he was freelancing conversion work and needed references. Monetization was weak. The pattern that stuck was different: find a sharp job people already pay for, ship a thin product, tell the truth in public, and let distribution compound.

Headlime was the first real AI score. He had headline formulas in a book. He turned them into software. Then GPT-3 arrived and he got into an early access cohort, one of the first people wiring that model into a production copy tool alongside the early Jasper and Copy.ai wave. Public accounts of the trajectory put Headlime at roughly $1K MRR in December 2020, about $20K MRR by February 2021, and a seven-figure sale to the Jasper orbit not long after. Eight months, more or less, from product to exit.

That is the clean legend. The next chapter is the one operators should actually study.

The moment something worked, and then did not

When Stable Diffusion hit, Postma moved again. ProfilePicture.AI went out in about thirty hours. Six-figure sales in a week, per his own public telling and the interviews that followed. Fun avatars. Feed fuel. Then the curve bent. Novelty faded. Lensa-style apps ate attention. Sales cooled.

Hold on. Most founders would either double down on virality or quit the category. He listened to what people were actually asking for in the replies and support threads: not a party trick profile pic. A professional headshot. The thing you need for LinkedIn when you will not book a photographer.

He launched HeadshotPro (and, in parallel, Deep Agency, an AI "modeling agency" experiment). Deep Agency got press. HeadshotPro got the money. That split is the whole piece if you are paying attention. Hype and cash are different instruments.

The boring product

HeadshotPro is intentionally narrow. Professional headshots. Not dating packs, not meme styles, not every horizontal the competitors ship to chase App Store screenshots. Pricing, as listed publicly for long stretches, has lived in one-time packages roughly in the $29 / $39 / $59 band for batches of images, not a pure seat subscription. Refunds for unhappy customers are part of the trust pitch. The site claims on the order of hundreds of thousands of customers and tens of millions of headshots generated. Treat marketing totals as marketing totals. The direction is still clear: this is a volume utility, not a toy.

The quality claim he has repeated in interviews is the opposite of "we wrapped open source and shipped." Custom models, long pipeline work, a year of tuning toward photos that do not immediately read as synthetic. Whether you buy the "indistinguishable" language or not, the strategy is coherent: win on the job people are embarrassed to get wrong, not on feature count.

The growth engine nobody screenshots

Launch week is not why HeadshotPro kept showing up.

SEO is. Indie Hackers writeups of his approach describe programmatic pages for city and long-tail headshot queries ("professional headshots in [city]"), plus ordinary blog content, aimed at keywords with enough volume and low enough difficulty to rank without a content army. Product Hunt launches were framed, in his own comments, less as pure user acquisition and more as backlink machines: even a mediocre ranking still throws domain authority around the web. When other apps go viral for AI headshots, search interest spikes, and the site already sitting on the term catches the overflow. He has described weeks where a competitor's virality tripled his sales because he owned the search intent.

That is not a 24-hour story. That is a six-month compounding story. It also means the copyable lesson is closer to "pick a monetizable search job and build plumbing for it" than "tweet a demo."

Pieter Levels retweeting early posts is part of the lore. It is also not a strategy you can purchase. What you can copy is the public logging, the willingness to kill the press-magnet product when the sales-magnet product is obvious, and the patience for SEO timelines that do not fit a hackathon video.

The asterisks

Here is where I get careful, because the internet is not.

The "$300K MRR" figure appears constantly in secondary writeups. HeadshotPro's public pricing has often been one-time packages, not classic monthly seats, and Postma has talked in later commentary about subscription mechanics being a trap for certain product shapes. Some databases show much smaller annualized snapshots for single years. Build-in-public numbers move. Portfolio revenue gets collapsed into one headline. I am not going to pretend a blogger's roundup is an audited P&L. What I trust more is the pattern across years: early GPT product sold for seven figures, AI image products that repeatedly printed meaningful cash, a holding company (Postcrafts) still owning HeadshotPro rather than flipping it on the first offer, and a slow admission that hiring help was necessary even for someone who marketed the solo myth.

He has said he should have built a team earlier. That sentence is worth more than another revenue screenshot. The lonely version of the empire does not scale quality pipelines forever. Compute bills do not care about your personal brand.

Competition is permanent. Every model release births ten headshot clones. His answer has been focus and distribution infrastructure, not feature parity with every clone.

The verdict

If you only remember the exit, you will try to time the next GPT wrapper. If you only remember the $300K line, you will argue about someone else's spreadsheet.

The useful story is narrower. Postma keeps winning the same way: enter a technology wave early enough to matter, ship something people already budget for, follow sales rather than press, and install boring acquisition (SEO, refunds, clear packaging) that still works when the feed is bored. HeadshotPro is not clever because it is AI. It is clever because professional headshots were a real purchase before diffusion models existed, and he made the purchase faster and cheaper without turning the product into a carnival.

The product was boring on purpose. That is the part worth stealing.

Sources

- Danny Postma on X

Your Brand Is Probably Wrong in ChatGPT. Here Is the 90-Minute Audit.

Agnel Nieves — Sat, 01 Aug 2026 14:46:33 +0000

This post is about auditing how answer engines currently represent your brand. It is not about rewriting your entire site in one weekend. By the end you will have a one-page gap matrix and a fix order you can hand to a founder without a 40-slide deck.

I keep meeting teams who shipped schema, refreshed the homepage, and still have no idea what ChatGPT says when a buyer asks "what does [company] do" or "best [category] tools for [job]."

You cannot fix what you have not queried.

What you need before you start

90 uninterrupted minutes
A spreadsheet with five columns: Query, Engine, What it said, Gap type, Fix owner
Four engines if you can: ChatGPT (with search/browsing if available), Claude, Perplexity, Google AI Overview or the AI Mode you actually see in your market
A human who knows the product well enough to spot a wrong pricing tier

Incognito or a clean workspace helps. Your logged-in personalization is not your buyer's world.

The query set (run all of them)

Numbered on purpose. Do these in order.

"What is [Brand]?"
"What does [Brand] do?"
"Who is [Brand] for?"
"How much does [Brand] cost?"
"Best [category] tools for [primary job]"
"[Brand] vs [Competitor A]"
"[Brand] vs [Competitor B]"
"Alternatives to [Brand]"
"Is [Brand] legit / trustworthy?"
"Recent news about [Brand]"
"[Primary product] tutorial" or "how to get started with [Brand]"
A category question that should include you but never says your name out loud, phrased the way a buyer would

If you sell in multiple segments, duplicate 5 and 12 per segment. Do not improvise mid-audit. Improvisation is how you cherry-pick good answers.

The gap matrix

For each answer, tag one primary gap:

Tag	Meaning	Typical fix
Missing	You should appear; you do not	Category proof, third-party mentions, clearer owned definition
Wrong	Invented feature, bad category, dead product name	Correct owned canon + earned corrections
Outdated	True last year, false now	Dated changelog, pressable update, refresh partners
Competitor-owned	Competitor is the default example in your category	Comparison pages are not enough; need independent cites
OK	Accurate enough to ship	Monitor quarterly

Caption for your screenshot folder: query, engine, date, gap tag. The caption is the section summary when you report up.

What the results usually look like

Patterns from client audits, not a claim about your brand specifically:

Query 1–3 are often "OK-ish" if you have any web presence. Soft wrong more than missing.
Query 4 (pricing) is a graveyard of invented tiers and "freemium" guesses.
Query 5 and 12 are where you discover you are not in the category set at all.
Query 6–8 are where competitors with better earned media win even if your product is stronger.
Query 9–10 surface old scandals, wrong founders, or total silence.

When the model is wrong, resist the urge to only update your homepage H1. Remember the citation mix: a large share of what models lean on is earned media, not your domain. Pair owned canon with something a third party can cite. See 84% of AI Citations Are Earned Media.

Fix order (after the 90 minutes)

Write the canonical 120-word answer to "What is [Brand]?" on a stable URL. Plain language. No banned marketing sludge. Date it.
Kill the most dangerous wrong (usually pricing or security claims) with a clear owned page and, if needed, a support note to partners repeating the old line.
Pick one category query you lost and build proof: customer artifact, benchmark with method, or independent review pitch. Not five thin blog posts.
Schedule a re-run of the same twelve queries in 30 days. Same order. Same engines.

Never start with a full rebrand because Claude misordered your feature list. Never start with a full rebrand because Claude misordered your feature list. Fix the fact, then the distribution of the fact.

Tiny experiment

Tomorrow, run only queries 1, 4, and 5 on two engines. Put the answers in a doc without commentary. Send them to your founder. Ask: "Would you fund a fix?" If the answer is no, either the gaps are small or the company is not serious about AI discovery yet. Both are useful information.

The scaffold for this audit is stable even when models change. The dial you turn is which gaps you fund first. The fail-state is screenshotting one flattering answer and calling the brand "AI optimized."

Sources

84% of AI Citations Are Earned Media (Promptway)
Muck Rack Generative Pulse

- Getting Your Writing Seen Beyond Your Own Site (Promptway)

I Made Grok 4.5 My Default Coding Model for One Client Week

Agnel Nieves — Sat, 01 Aug 2026 14:46:32 +0000

On July 8, Cursor and SpaceXAI put Grok 4.5 in front of every Cursor subscriber who would click the model picker. The pitch was simple: frontier-ish coding and agent work, trained with real Cursor interaction data, cheap enough that token anxiety stops being the main character.

I did what I always do when a tool claims it belongs in the daily path. I made it the default for one week on a real client repo and wrote down what broke.

This is not a benchmark re-score. It is a work log.

The task I actually ran

Client context, sanitized: a Next.js App Router publication with content validation, MCP servers for ads and analytics reads, and a pre-commit asset pipeline. The week had three jobs that match how I actually bill time.

Ship a feature branch with two new MDX routes, shared components, and schema fields that had to pass bun run validate.
Debug a flaky integration where an MCP tool returned partial JSON and the agent kept "fixing" it by inventing keys.
Drive a multi-file refactor that touched route handlers, a Zod schema, and a small TypeScript helper without breaking static generation.

I used Cursor as the shell the whole week. Grok 4.5 as the default model. Same MCP config I already trust. No special harness beyond what a mid-size agency repo already has.

What worked, with receipts

Speed on the boring middle. Autocomplete-plus-edit on component props, frontmatter keys, and test stubs felt fast. Not magically smarter than Opus-class on hard architecture. Noticeably cheaper per loop when I was iterating ten times on the same file. For the kind of work that is 70% of a billable day, that matters more than a one-point leaderboard bump.

Cursor-native habits. This is the underrated part of "trained with Cursor data." The model was less confused by multi-cursor edits, partial selections, and "apply this diff but keep my comment." That is not a general intelligence claim. It is a product claim, and on this repo it held.

Price as a product feature. At roughly $2 input / $6 output per million tokens (and lower on cached), I stopped doing the mental math that makes people under-prompt. I asked for the second rewrite. I asked for the third. The quality of the session went up because I stopped rationing turns. That is a real effect, even when single-shot quality is a wash.

Rough week spend on the client branch, all-in model cost through Cursor: under what a single heavy Claude Max-style day used to burn on comparable volume. Your numbers will differ. The direction did not.

What broke first

Unsupervised multi-file refactors. When I pointed it at "migrate these three modules and keep generateStaticParams honest," it produced a plausible plan and then a half-applied migration. Two files updated. One left mid-state. Tests red. Claude Code on the same prompt the next morning finished the graph more cleanly, with fewer "I will fix the types in a follow-up" lies.

Confident wrong tools. On the MCP JSON bug, Grok 4.5 was eager. It rewrote the client twice before it agreed to log the raw payload. Claude's slower, slightly pedantic loop would have asked for the log earlier. Eager is not free. Eager costs review time.

Long context discipline. On a chat that had already eaten a big schema dump, it started compressing my constraints into vibes. I had to re-paste the Zod refine rules mid-thread. Not unique to Grok. Worse here than my Sonnet baseline on the same thread length.

How it compares on the jobs that matter

Job	Grok 4.5 in Cursor	My Claude Code / Sonnet baseline
Tight UI and content schema edits	Winner on speed + cost	Fine, pricier per loop
MCP / tool debugging	Mixed; needs tighter human steering	Winner on "slow down and inspect"
Multi-file refactors with static guarantees	Risky without babysitting	Winner for unsupervised depth
"Just ship the PR" Friday afternoon	Strong if you stay in the editor	Strong if you live in the terminal agent

The posture split from the public discourse is real. Cursor assumes you are editing. Claude Code assumes you are delegating. Grok 4.5 inherits the Cursor posture even when you ask it to act like an agent. That is not a bug if you wanted a co-pilot. It is a bug if you wanted a night-shift senior.

Verdict

Buy Grok 4.5 as a default inside Cursor for iterative product work if you already live in that editor and your pain is token cost plus turn latency.

Wait before making it the only model for unattended multi-file agents or production refactors that touch codegen boundaries. Keep a heavier model one hotkey away.

Skip if your whole workflow is terminal-agent and you already have Claude Code tuned. Switching shells just to chase a model is a tax.

I am not deleting Claude from the stack. I am not writing a "Grok won coding forever" post. I am changing my default for the 70% path and keeping the expensive brain for the 30% path that ships the scary PR.

Six months ago I would have told you to pick one model and learn it deeply. I no longer believe that is the right advice for coding agents. The July wave made routing the skill. Grok 4.5 is a strong cheap route, not a religion.

What I will re-test in 30 days

Same multi-file refactor with whatever patch they ship after the first wave of forum complaints.
Token spend on a pure greenfield app versus this brownfield content site.
Whether "double usage first week" promo pricing was hiding a bill shock at steady state.

If any of those flip the verdict, I will update this page in public. That is the deal on The Stack.

Sources

- Prior Promptway stack note: Connecting Claude to Google Ads and GA4 via MCP

The July Model Wave Is Not a Race You Need to Win

Agnel Nieves — Fri, 31 Jul 2026 15:39:17 +0000

Three frontier launches. Two weeks. One bad habit.

The habit is crowning a winner from a press release. Claude Sonnet 5 on June 30. OpenAI's GPT-5.6 family rolling into general availability around July 9. Grok 4.5 on July 8, co-trained with Cursor and priced to make coding agents feel cheap. The charts moved. The posts multiplied. The claim underneath most of them was the same: this is the model you should standardize on.

[The claim is nonsense. Standardization is the risk. Routing is the skill.]

What actually shipped

Strip the demos. Keep the operator facts.

Model	Maker	Window	Operator-relevant shape
Claude Sonnet 5	Anthropic	late June	Balanced agent runs, coding, long reliable chains
GPT-5.6 Sol / Terra / Luna	OpenAI	late June to mid-July	Tiered family: flagship Sol, everyday Terra, cheap Luna
Grok 4.5	xAI + Cursor	July 8	Coding and agent work at aggressive API pricing

OpenAI gated GPT-5.6 longer than the others. Safety review, staged partners, then broader access. That is part of the product story now, not a footnote. Anthropic and xAI moved faster to availability. Access policy is a feature.

Open source did not wait. GLM-5.2, DeepSeek V4, Qwen 3.6 and peers kept closing the gap for hosted and self-hosted work. The frontier is crowded. The "one brain for everything" era is over as an architecture choice, even if the marketing still pretends otherwise.

Ranked by Tuesday impact, not leaderboard theater

1. Cost and tiering matter more than the top score.
OpenAI shipping Luna / Terra / Sol as a family is the real product decision. You can route a triage job to a cheap tier and a hard research job to a flagship without changing vendors. That is operator infrastructure. A single "best model" headline is not.

2. Grok 4.5 inside Cursor changes the default coding bill.
A model trained with Cursor interaction data, sold at roughly $2 / $6 per million tokens, is not a vibe. It is a budget line. Teams that were bleeding token spend on heavier agents will try it this month whether or not they rewrite their stack. Watch adoption, not the CursorBench slide.

3. Sonnet 5's value is reliability under load, not a new personality.
If your workflows are multi-step agents that must finish, a mid-tier that holds the chain beats a flashy flagship that drifts. That is a boring metric. It is also the one that shows up in support tickets.

4. Export-control residue is still on the board.
Anthropic's Fable 5 / Mythos-class redeploy after the June export-control pause is a policy story wearing a model name. If your product depends on a single frontier weight class, you now have a quarterly risk review whether you wanted one or not.

5. Benchmarks withheld or partial are a signal.
When a lab ships without the usual suite, read that as intentional. Not always sinister. Always incomplete. Do not fill the blanks with Twitter confidence.

The permanent-winner myth

Here is the strongest form of the bad advice, stated fairly:

"Pick the best model now, standardize the company on it, and stop thrashing."

Sounds like discipline. In a market that ships a capable model roughly every few days once you count open weights, it is how you bake technical debt into the org chart. Hard-wiring one provider turns every release into a migration project. Abstracting the model turns every release into a config change and an afternoon eval.

I will say it once. Build so you can swap. Test on your tasks. Route by job. The July wave did not crown a champion. It confirmed that several frontier options are close enough that your data decides the winner, and that the winner can change by workload.

What to do this week

Not a tutorial. A short list with falsifiable edges.

Write down the three jobs AI actually does in your stack. Not aspirational. Actual.
Run each job on two models you already pay for. Same prompt. Same inputs. Score quality, latency, and unit cost.
Kill one permanent default if it loses on two of three metrics. Replace it with an explicit route.
Re-run the same three jobs when the next major weight lands. Calendar it. Do not wait for vibes.

If you skip this, you will still "evaluate models." You will just do it in Slack arguments instead of on paper.

What I am grading later

Three calls. Score me in September.

Teams that abstract the model layer will switch defaults at least once before Labor Day without a rewrite. Falsifiable: count public postmortems and internal changelogs that mention a one-line model swap.
Grok 4.5 will take measurable coding-agent share from pure Claude/OpenAI defaults among Cursor-heavy shops. Falsifiable: usage surveys, spend reports, forum default chatter.
The "one model for the company" memo will keep getting written, and it will keep aging poorly within 90 days. Falsifiable: watch the memos, then the quiet exceptions.

Noise to leave on the floor

Affiliate "best model of July" roundups with forty logos. Demo videos that never show a failure case. Seat-count panic recycled from February without a single cancelled contract attached. Benchmark charts with no task definition.

The July wave is real. The race framing is residue.

Build for the next release, not the last press cycle.

Sources

- FelloAI, Best AI models in July 2026

Six Months After the SaaSpocalypse, the Stocks Came Back. The Operator Story Did Not.

Agnel Nieves — Fri, 31 Jul 2026 15:39:16 +0000

Quote the panic first. In the strongest form.

Anthropic shipped agentic tools that do real office work. If one AI can replace five SaaS seats, the entire software multiple is wrong. Hundreds of billions in market cap should vanish, and they did.

That was the February story, compressed. Claude Cowork, plugins, Opus-class upgrades, a Goldman-style software basket down hard in a session, estimates of roughly $285 to $300 billion erased in the worst prints. Workday and Salesforce and a long list of names ate the tape. The word SaaSpocalypse did the rest.

[The market move was real. The monocausal story was not. The operator aftermath is still the open question.]

Claim Chowder: February versus July

Claim (Feb 2026)	July grade	One-line reason
"Agents kill SaaS this year"	Fail	Public multiples mean-reverted faster than seat bases dissolved
"Cowork is only a demo"	Fail	Workflow automation is now a default expectation in buyer RFPs
"Every horizontal SaaS is dead"	Nonsense	Data gravity, compliance, and workflow lock still pay rent
"AI is a pure margin tax on software"	Partial	Pricing pressure is real; free replacement is not universal
"Operators will cancel seats in bulk by summer"	Incomplete	Spot cancellations yes; systematic wipeout not in the open data

I am not here to defend software equities. I am here to stop operators from taking a Wall Street weekend as a strategy memo.

What the selloff actually was

Three things collided.

A visible agent product that looked like it could draft, file, research, and chain office tasks without five logins.
A valuation regime that had already been compressing. Software multiples had room to fall before Cowork gave traders a narrative.
A meme with a deadline. Once "SaaSpocalypse" fit in a headline, every downtick confirmed the story and every uptick was "denial."

Executives called the panic overblown. Analysts warned about pricing power anyway. Both can be true. A 6% session in a sector basket is not a customer survey. It is a forced rewrite of discounted cash flow assumptions under uncertainty.

What operators actually did

From the ground, not the tape:

Trials exploded. Migrations lagged. Trying Cowork or a coding agent is cheap. Ripping out the system of record is not.
Shadow AI spend rose next to SaaS spend. The line item grew. It did not always replace. CFOs noticed both.
Categories split. Thin workflow glue and generic document tools felt heat. Vertical systems with proprietary data, audit trails, and integrations held. Boring infrastructure is still boring infrastructure.
The seat is not dead. The idle seat is. Buyers started asking which licenses produce artifacts an agent could produce by Tuesday. That question does not require a stock chart.

If your product's only defense is "we have a nicer UI for a task Claude can already do," you felt this winter. If your product is the system of record for something regulators care about, you felt a pricing conversation, not an extinction event.

The residue that matters more than the rebound

Stocks can rebound on rates, earnings, and attention cycles. Three structural shifts did not rebound away.

1. Buyers budget for agents as a category.
Not experimental. Line-item. That permanently changes how they negotiate your renewal.

2. Switching cost is the product.
Feature checklists lost. Data model, integrations, permissions, and eval history won. The SaaSpocalypse was a referendum on shallow SaaS.

3. Narrative risk is now product risk.
A lab demo can reprice your category in a session. That is not fair. It is the water you swim in. Communication and proof-of-work matter more when the market is itchy.

Falsifiable calls for Q4 2026

Score these in December.

Net seat reductions in generic productivity SaaS will show up in earnings language more than in February price charts. Look for "optimization," "consolidation," and "AI substitution" in prepared remarks, not just stock ticks.
Vertical SaaS with proprietary workflows will guide stable or rising NRR while horizontal doc tools guide down. Split the basket; stop averaging apples and glue.
At least two public software companies will rebundle around agent runtimes rather than human seats as the primary SKU. Watch packaging, not keynotes.
The next Cowork-class launch will move equities less than February did, and move RFPs more. Attention decays. Procurement language sticks.

If all four miss, I will say so in a follow-up. That is the job.

What to ignore for the rest of July

Death-of-SaaS threads with no cancelled contract attached. Victory-lap founder posts that equate a stock bounce with product safety. "We added AI" changelog spam that does not change the artifact the customer ships.

Six months later, the apocalypse looks like a repricing. The operator story looks like a filter. Shallow tools got priced like shallow tools. Deep systems still have to prove they are deep.

The chart healed. The question did not.

Sources

- Business Insider coverage of Claude Cowork and market reaction (Feb–Apr 2026 cycle)

Indie Hackers Built AI Factories. Then Nobody Showed Up.

Agnel Nieves — Thu, 30 Jul 2026 15:34:35 +0000

There is a particular kind of demo video making the rounds this year. Dark editor theme. Agent panel streaming. A full app appearing in minutes. The caption says some version of "the only limit is imagination."

Then you click through to the product and the waitlist is a Google Form with eleven responses, three of them the founder under different emails.

Pieter Levels put a sharper point on it in June: indie hackers building fancy AI factories with no money and no traffic. The post traveled because it was mean in the useful way. It named the thing people were privately embarrassed about.

I do not think the builders are stupid. I think the bottleneck moved, and a lot of playbooks did not.

The setup

For a decade the indie story had a honest hard part: shipping. You had to learn enough code, or hire it, or glue no-code until it cried. AI coding agents did not remove craft. They removed the excuse. Claude Code, Cursor, Grok-class models in the editor, boilerplate kits like ShipFast, component generators, database scaffolds. The factory floor got automated.

What did not get automated: a reason for a stranger to care, a channel that compounds, a promise people renew.

So we got a generation of products that are technically complete and socially imaginary.

What the empty factory looks like

I keep a private list from conversations and public postmortems. Patterns, not call-outs.

Beautiful onboarding, no top-of-funnel. The first-run experience is gorgeous. The homepage has no distribution plan beyond "post on X."
Feature parity with a giant, trust parity with a stranger. "We are like Notion + AI" is not a wedge. It is a request to lose.
One-week revenue, zero week-eight retention. Launch spike from friends and a Product Hunt bounce. Then silence. Same shape as the flight-sim money that visits once and leaves. See Pieter Levels and the sim that went to zero.
Agent theater instead of a job. The product demos an agent. The customer still has a job to do on Tuesday that the agent does not finish.

Hold on. Is this just "marketing matters"? Yes, with a 2026 asterisk: marketing also got flooded with AI sludge, so the trust bar for small tools went up while the build bar went down. You are not competing with silence. You are competing with a thousand competent-looking clones.

Who still prints money

Look at the people this column has already covered and the durable ones like Danny Postma on HeadshotPro.

The pattern is rude and consistent.

A purchase that existed before the model. Headshots. Photo generation for a known vanity or work need. Boilerplates for people who already sell software. Verified revenue pages for people already lying with screenshots.
A channel that is not a single launch post. SEO, newsletters, years of public logging, communities where the founder is a known quantity.
Willingness to be boring. One vertical. One clear price. Refunds. Support. The unglamorous middle.

Marc Lou's 24-hour TrustMRR spike was real, and so was the year he earned less than the prior one. The weekend needs a decade under it. That is not romantic. It is accounting.

The nervous part

Here is the sentence founders flinch at: if your only advantage was that you could code the v1, you no longer have an advantage.

The new hard skills look like this:

Picking a job where "done" is obvious and payment is habitual
Building proof other humans will cite (see the earned-media AEO problem for brands; it applies to products too)
Saying no to agent features that do not finish the job
Measuring renewals harder than launch-day MRR screenshots

AI factories without trucks are still factories. They just manufacture unsold inventory faster.

Field notes, not a sermon

I asked operators what they cut when the empty dashboards got embarrassing. The answers rhymed.

Cut model roulette. Pick a stack. Ship weekly to real users.
Cut horizontal scope. One persona, one job.
Cut launch culture as identity. Replace with a channel experiment that lasts thirty days minimum.
Keep the agents, but point them at distribution chores (support macros, content diffs, SEO page generation for a keyword you already validated) instead of only at more product surface.

None of that is as fun as a three-hour build video. All of it is closer to Photo AI at high margin than a sim at $0/m.

The verdict

The vibe-coding era did not kill indie hacking. It killed the story that shipping was the main boss fight. Levels' jab lands because the feed is full of factories and short on customers. The founders who will still be here in 2027 are not the ones with the most tools in the agent panel. They are the ones who can explain, in one sentence, who pays them every month and why that person would be annoyed if the product disappeared.

Build the truck. Then build the factory.

Sources

Pieter Levels, indie hackers AI factories post
Pieter Levels flight sim story (Promptway)
Marc Lou 24-hour app (Promptway)
Danny Postma / HeadshotPro (Promptway)