I read every reply to every issue of this newsletter. I don't always respond quickly — the combination of writing, running experiments, and being a husband and father in my late fifties means some replies sit longer than they should before I get back to them. But I read them.

Over the past four weeks, a few questions have come up enough times, from enough different people, that they deserve a full answer in the newsletter rather than individual replies. Today I want to work through the three that have come up most.

"How do I know where to start if I don't know which stream to pick?"

This is the most common question, and it's the right first question.

My answer: start with the stream that matches your existing skills most directly, and treat it as a learning experience rather than a revenue strategy.

Here's the reasoning. When you're learning to use AI tools effectively — which is a real learning curve, whatever anyone tells you — it's significantly easier to learn on material you already understand. If you're a marketer, start with content. If you have technical skills, start with apps or automation. If you're a writer, start with writing-adjacent streams.

Starting with the stream that's furthest from your existing competence, just because it seems like it has higher upside, means you're learning AI tools and the subject matter simultaneously. Both are manageable alone. Together they can be overwhelming enough to produce paralysis.

Start where you're already strong. The AI gives you leverage on what you know.

"Isn't all of this going to get disrupted by better AI? What's the point of building streams that might be obsolete in two years?"

This comes in different forms — sometimes hostile, sometimes genuinely anxious — but the underlying question is the same: if AI keeps improving this fast, doesn't it eventually eliminate the income streams you're building?

My honest answer: yes, some of what we're building will need to evolve significantly as AI capabilities advance. That's not unique to AI-enabled income streams — it's true of any business in any era of technological change.

The practical question is whether the specific streams we're testing are more or less durable than alternatives. My assessment: streams built on real relationships, genuine content with a specific human voice, and community trust are more durable than streams built purely on production efficiency. The production-efficiency advantages of AI will eventually be available to everyone. The relationship and voice advantages require decades of a specific human life to build.

This is also why I'm not building a stream that depends on AI being better than humans at something. I'm building streams that use AI to extend what a specific human — me — can do. If AI gets better, I use the better AI. The human judgment layer doesn't become obsolete because the tools improved.

"How do I get my spouse/partner on board with this?"

I was not expecting this question as often as it arrived. But it's shown up enough times, with enough genuine vulnerability in the phrasing, that I want to address it seriously.

My wife asked me the same question I imagined everyone asking: "Is this real, or is this the next thing you're spending three months on before moving to the next thing?"

The honest answer I gave her, and that I'd give anyone asking this: it's real if you treat it like it's real. The difference between an experiment that produces results and an experiment that burns six months and produces nothing is usually not the quality of the idea — it's whether the person running it is willing to hold themselves to honest evaluation criteria and make decisions based on what the evidence shows rather than what they want it to show.

For the partner who's skeptical: set specific milestones with specific timelines. Not "we'll be profitable in year one" — that's too vague to evaluate. Specific: "by month six, the newsletter will be at X subscribers with Y engagement rate. If it's not, we reassess." Give them the criteria, and then meet them.

For the person asking their partner: show your work in public. This newsletter exists partly because doing things in public creates accountability that doing things privately doesn't. Your partner's skepticism is reasonable and is best answered by evidence, not persuasion.

An invitation.

These issues are better when the questions driving them are good. Reply to this with what you're working on, what you're stuck on, or what you want me to address in a future issue.

I can't promise I'll answer every reply individually, but I can promise the questions shape what I write. The newsletter is better for being a conversation than a broadcast.

This Week in AI: The question of AI's effect on relationships — including working partnerships and marriages where one person is "in the AI world" and the other is skeptical or disengaged — is starting to get serious attention from organizational psychologists and relationship researchers. The dynamic of one partner living in a rapidly changing technology environment while the other doesn't is genuinely novel, and the communication strategies that work for it are still being figured out.

Join the conversation with other builders in The Upstream community at start.tenstreamslab.com. Sometimes the answer you need is from someone who's two steps ahead of you, not from the newsletter.