Thoughts & Perspectives
Building in public
TRAIGA: What the Texas Responsible AI Governance Act Actually Requires of a Business (2026)
TRAIGA, the Texas Responsible AI Governance Act, took effect January 1, 2026. It applies to any business that operates in Texas, with no size exemption, but most of its duties bind state agencies. For a private company it is a short list of things you may not build AI to do on purpose, one disclosure duty for health care, and fines that start at $10,000. Here is the bill text in plain English.
Claude for Nonprofits (2026): The $8 Team Plan Seat, Who Qualifies, and What a 10-Person Team Actually Pays
Claude for Nonprofits, as of September 2026: Anthropic lists Claude Team at $8 per user per month for verified nonprofits, against $20 to $25 on the standard plan, and $3 for nonprofits in low- and middle-income countries. Minimum 2 seats. Under 20 people you verify through Goodstack in a few minutes; above that, it's a sales call. Who qualifies, what's included, and the math.
AI for Nonprofits (2026): What Every Vendor Actually Charges You, and the Three Jobs Worth Doing First
AI for nonprofits, as of September 2026: Claude and ChatGPT both land at $8 a seat for verified nonprofits, Google Workspace Business Standard is $3.50, and Claude adds a $3 rate for lower-income countries. Every vendor verifies you through the same kind of check. The price table, what to do first (grants, reporting, member and donor communication), and the policy to write before anyone pastes donor data into a chatbot.
Claude Opus 5.5 (2026): What It Costs, What Changed From Opus 5, and Whether Your Business Should Switch
Claude Opus 5.5, as of September 2026: announced September 22, $4 per million input tokens and $20 output (Opus 5 was $5 / $25), with cache reads cut 60% to $0.20. Anthropic says it runs at Fable 5.1's level for 40% less than Opus 5. The price sheet, the workload math, what happened to your usage limits, and who should switch.
AI Grant Writing (2026): What Funders Actually Allow, a Workflow That Keeps You Honest, and a Copy-Paste Prompt
AI grant writing done responsibly, as of September 2026: what funders say on their own pages (NIH won't treat applications substantially developed by AI as original work; NSF encourages disclosure and holds you to accuracy), a seven-step workflow where AI organizes and critiques while you supply the facts and the story, and a copy-paste prompt that scores a draft against the funder's criteria.
AI for Associations (2026): What It Does for Member Services, Events, Renewals and Certification, and What Stays Human
AI for associations and member organizations, as of September 2026: the six jobs where it earns its keep (member services, events, content, renewals, certification and education, governance), the three things that stay human (exam content, member data, dues and discipline decisions), and why a 501(c)(6) shouldn't assume it gets the nonprofit discount.
Claude Plugins for Business (2026): What They Are and Where They Run
Claude plugins bundle skills, connectors and sub-agents into one install, on every paid Claude plan. As of September 2026 there's a directory with a developer submission portal (September 25) and a Salesforce plugin in beta (September 15). What a plugin is, where it runs, what it costs, and what an owner can actually do with one this week.
AI Governance Framework (2026): What NIST and the EU AI Act Actually Ask of a Business Your Size
An AI governance framework is the set of rules for how your company picks, uses and checks AI: who owns it, what data goes in, and who reviews what comes out. As of September 2026 the reference framework in the US is NIST's AI RMF (voluntary, four functions, now being revised), and the EU AI Act has applied since August 2, 2026. Here is what each asks, and the version a 10 to 200 person company can run.
SuperGrok Pricing (2026): Lite $10, SuperGrok $30, Plus $100, Heavy $300
SuperGrok costs $10 a month for Lite, $30 for SuperGrok, $100 for Plus and $300 for Heavy as of September 2026, or $100 to $3,000 a year billed annually. Business is $30 a seat. x.ai's main pricing page still leaves Lite and Heavy blank; its checkout page prints both. Here is every tier, what each adds, which ones include Grok Bot, and which one to buy.
Grok Bot (2026): What It Is, What It Costs, and Whether It's Worth It
Grok Bot is x.ai's always-on AI teammate: agents with their own cloud computer that sign into your tools and finish jobs while you are offline. As of September 2026 there is no standalone price. It comes with SuperGrok ($30), SuperGrok Plus ($100), SuperGrok Heavy ($300) and every paid Cursor plan from $20, each with a weekly allowance and optional pay-as-you-go after it.
Claude Science (2026): What It Is, Who Gets It Free, and What Stays on Your Machine
Claude Science is Anthropic's beta workbench app for researchers, not a new model. It runs analyses on your own laptop or cluster, comes with Pro and up, and verified academic labs can get Team seats at no cost to start. Here is what it does, what it costs, and what actually leaves your machine, as of September 2026.
Fractional CIO (2026): What It Costs, vCIO vs Fractional, and Who Should Actually Hire One
A fractional CIO is a senior IT leader you rent part-time to own the systems, vendors, security and now the AI decisions, instead of hiring a CIO at a $175,140 BLS median plus $13,399 of employer payroll tax. The same seat is sold under another name, vCIO, usually inside a managed-IT contract, and that difference matters more than the price.
Ready to scale your product vision?
I help high-growth startups bridge the gap between technical complexity and business revenue.
Book a Strategic Call