I Had 535,000 Followers. Why Was Nobody Seeing Me?

In 2009 I cracked the code on  how to grow an audience on Twitter. Followers were everything then when the social media algorithm  was the social graph. That meant that if people followed you, they saw what you posted, so I spent years building a following and working with the system. And I had a significant 535,000 followers. That was the old game.

But 17 years later, on June 24, 2026, my posts were only seen about 200 times. The social media term for that is “impressions”. The reality is that compared to 2014, I was almost invisible. The new interest graph pioneered by TikTok had changed the game.

The gap between the old and the  new was vast and it meant that my investment of time and money back then was now worth almost zero.

The gap

That gap is the story and the history of social media. 

The platforms have moved from the social graph, where the people who follow you decide who sees you, to the interest graph, where the algorithm decides what each person is interested in and shows them that. 

Your followers are now one signal among many. And the feed you’re competing in is filled with content that AI can produce in seconds, at almost no cost, without end.

The experiment

So in June I started an experiment. 

I’m 69, and I wanted to know whether I could crack the code a second time: how do you get seen on the interest graph, in a flood of AI content? I posted every day and used AI to analyse every number X gave me.

Four months later, my posts have been seen 2.7 million times since 1 July, and I’ve gained more than 3,000 followers. In October I averaged 35,000 views a day. 

When I started, it was 200.

But views don’t tell you who is listening. So this week I went back through every reply to my posts since July: 3,660 replies from 2,020 people. I used AI to help me sort and count them.

Followers no longer decide who sees you

The clearest sign is in my own replies. Only about one in five of the people who replied to me were following me. If they’re any guide, most of the people seeing my posts weren’t on my follower list at all. X “had” decided that they’d be interested.

To be fair, the algorithm isn’t hiding me from every old follower. Many accounts from 2009 to 2015 are no longer active. But even allowing for that, a typical post today reaches less than 1% of my followers. The audience I spent years building now buys me very little reach.

So what does get you seen? So using X as a platform to do some research from the 3,660 replies that had responded to my posts on X. And these are the insights that have changed how I create content and how I use social media on X. 

1. Name the feeling

On the interest graph, the algorithm watches what people stop for, and shows the post to more people like them. So the question is what makes a stranger stop.

Since early October I’ve posted ten questions in the same format. Four of them took off, reaching between 12,600 and 25,100 people. The other six stalled below 3,600.

The difference wasn’t the time I posted or the length. Every question that took off asked how AI makes you feel. “Does AI make you feel left behind, or let back in?” drew 283 replies. “Has AI made you more excited about your future, or more uncertain?” drew 313.

The ones that stalled asked practical questions, such as whether starting again is harder now than at 30, or who still asks for your advice. They were fair questions. But they asked people to think, not to feel.

People don’t stop for a topic. They stop for a feeling they recognise in themselves. A machine can produce endless content about AI. It’s much harder to fake the moment someone reads a question and thinks, “that’s me.”

Even so, only about one in three of those questions broke through. The answer isn’t to post more. It’s to take more shots with the kind of post that can land.

2. Write for an audience, not a topic

The interest graph doesn’t only decide how many people see you. It decides which people. Write broadly and it will find you a broad crowd.

From July to September I wrote about AI in general: its risks, its paradoxes, what it does to thinking. Those posts drew a crowd, but the wrong one. About four in ten of the real people replying were debating AI itself. Very few were thinking about their own next chapter.

In October I narrowed the lane. I wrote for experienced people working out what AI means for their own work and life. Since mid-October, nine in ten of the real people replying fit the four groups I’m building Zyrro for:

• Builders making something new: 30%
• Rediscoverers getting back skills and energy they thought they’d lost: 22%
• Defenders of experience, arguing that judgment matters more now, not less: 20%
• People who feel threatened or worn out by the pace of change: 18%

A community is starting to form. About half of the people replying since mid-October had replied before, and 232 people have now replied across more than one period. Compare that with the July to September crowd: only 98 of 582 came back in October.

The posts that brought the most followers were personal, not clever. A post about why I’m not retiring brought 304 profile visits and 41 new followers in a day. The story of the year my business collapsed and my marriage ended brought 524 profile visits, more than any post since July. Personal detail tells the algorithm, and the reader, exactly who you are for.

3. Count people, not replies

This was a surprise. The flood of AI content isn’t only in the feed. It’s in the replies.

As I read, I kept seeing the same kind of reply: polished, agreeable, and saying back to me what I’d just said in slightly grander words. “The good question is the hard bit, the agent will happily answer a bad one.” One account had replied to me 72 times. Its bio said it was building a tool that writes replies for people.

So I counted. In July to September, about four in ten replies to my posts read as machine-written. Since mid-October it’s about one in four, from 139 different accounts.

This isn’t just my account. Imperva’s 2025 Bad Bot Report found that bots made up 51% of all web traffic in 2024. The machines aren’t only reading the internet now. They’re joining the conversation.

The effect is easy to miss. Reply counts no longer measure connection. If you judge your content by its replies, you are partly judging how attractive it is to software. I now count people: profile visits, new followers, and real people who come back.

4. Reply more, don’t post more

When anyone can publish endless content, adding more of your own is the weakest move. I learned that the hard way.

In the week of 20 July I posted 22 times a day, and at first it worked. Views jumped from about 7,000 a day to 55,000. So I kept going at close to that pace into August. Reach fell week after week.

In October I posted about 8 times a day. Views were higher than in August, and profile visits per 1,000 views tripled, from 1.0 to 3.1.

When I lined up the daily numbers, profile visits barely moved with how many posts I published. They moved with how many replies I wrote to the people in my threads. Posting is the shop window. Replying is the conversation that brings someone inside. And a real conversation is the one thing the flood can’t copy at scale.

There’s a cost to this, and I should be honest about it. I did all of it by hand, and replying became a treadmill. The next stage has to fit into about an hour a day.

5. The people reading aren’t always the people replying

The replies told me who was talking. X’s own audience data told me who is listening, and they are not quite the same people.

The people who click through to my profile are older than the people who reply. Over the past week, 68% of them were 45 or older, and one in four was over 65. Four in ten of the people who reply are between 25 and 44.

That gap matters. The quiet readers read, think and then visit my profile to decide whether I’m worth following. Most of them never reply to anything. If I judged my work only by the replies, I’d end up writing for the wrong people.

The bios of the people who do reply tell the rest of the story. One in four works in tech or software, one in five is focused on AI and one in five runs a business. Most are not influencers: 61% have fewer than 500 followers. These are working people trying to work out what comes next.

What this means if you’re trying to be seen

If you built an audience the old way, those followers aren’t wasted. But they no longer guarantee that anyone sees you. Each post now has to earn its reach from strangers.

You don’t need a big following for any of this to apply. Most of the people replying to me have fewer than 500 followers. The lessons work at any size.

• Stop counting followers. Watch profile visits and new followers per post instead. They tell you whether strangers liked what they found.
• Write to a group of people, not about a subject. “AI” is a subject. “Experienced people working out what AI means for their work” is an audience. The second one tells you what to write next.
• Name the feeling. Ask how something makes people feel, and they’ll answer from their own life. Ask what they think, and you’ll get a debate.
• Tell your own story, odd details included. The posts that earned followers were personal and specific: a year, a loss, a decision. Polish is cheap now. Detail is not.
• Count people, not replies. Likes and reply counts now include a lot of software. Profile visits and new followers are harder to fake.
• Spend your time on the people who come back. Replying to the same real people again and again is how the following becomes a community.

One caution. 

This is one account, on one platform, over four months. X decides who sees each post, and it can change the rules tomorrow. That’s why every win on X has to end somewhere I own: an email address, a newsletter subscription or a waitlist.

What I’m doing next

The experiment showed that visibility can come back, even after years of shrinking reach and even in a flooded feed. Now it needs to turn into something lasting. Here is the plan for the next month:

• Two feeling questions a day, and a record of which ones break through.
• About an hour a day of replying, spent on real people, especially the ones who keep coming back.
• A weekly scorecard with five numbers: views, profile visits, new followers, real replies and sign-ups to the Zyrro waitlist.

That last number matters most. And the reason why is the real lesson of these four months.

What 535,000 followers couldn’t tell me

In 2009 I was 52 and starting again after my business collapsed. I got up at 4:30 every morning to write before the day began. Every new follower felt like proof that someone out there was listening.

So I chased the number. For years it went up, and I took that as a sign I was being heard. I never asked who those people were, or why they came, or whether they were still there.

A follower is someone who once clicked a button. That’s all. The number can’t tell you if they still read you, or what they need, or whether you’ve ever helped them.

The 2,020 people who replied this year told me far more. So did the quiet readers who never replied at all. One in four of the people who visited my profile last week was over 65. My guess is that many of them are where I was in 2009: good at something, unsure what comes next, and wondering if it’s too late to start.

That changed what I think being seen means. In a feed full of machine-made content, being seen isn’t about reach. It’s about being recognised by the right people for the right thing. And nobody can recognise you for something you haven’t worked out yourself.

My reach didn’t come back when I found a better posting time or a cleverer hook. It came back when I decided who I was writing for and what I wanted to be known for. Everything else followed from that.

That’s the hard part. Not the algorithm. Not the schedule. Knowing what you want to be known for next.

Most people with long careers never get asked that question. They get asked what they do, or when they’ll retire. Nobody asks what they want the next 10 or 20 years to stand for.

That’s why I’m building Zyrro.

It’s an AI tool for people with long careers who aren’t sure what comes next. You share the stories of your working life. It finds the strengths running through them, shows you four possible paths and gives you a small first step to take this week.

When I started this experiment in June, I wanted to know one thing. Could I still be seen?

Four months and 3,660 replies later, I’ve found a better question. Seen by whom, and for what?

535,000 followers taught me how to be counted. 2,020 people taught me how to be found.

So here’s my question for you. If the right people found you tomorrow, what would you want them to find?

Join the Zyrro waitlist

You’ve spent decades building experience. Zyrro helps you work out what to do with it next.

Tell it the stories of your working life. It will show you the strengths you keep coming back to, four paths worth testing, and one small step to take this week.

We’re opening Zyrro first to a small group of early testers. If you’ve read this far and thought “that’s me,” I’d like you to be one of them.

Join the Zyrro waitlist →

This is how I did it

The numbers: X Analytics exports for my account, July to October 2026, plus X’s audience breakdown for the past week.

The replies: every reply to my posts from 7 July to 22 October 2026, read along with each replier’s public bio. Each reply was sorted into a group by what it said. For July to September, the shares come from a random sample of 360 of 1,408 replies.

“Real” replies means replies from people, leaving out machine-written replies, one-word replies and sign-ups to a test I ran.

The limits: this is one account on one platform. Only 127 people stated their age, and X’s age figures are its own estimates.

The help: I used Claude to collect, sort and chart the data. The conclusions are mine to defend.

The post I Had 535,000 Followers. Why Was Nobody Seeing Me? appeared first on jeffbullas.com.

* This article was originally published here

Start making $100+ per day this week with the best dfy system - Subscribe here!

What 733 People Over 50 Told Me About AI (Original Research)

Executive summary

In one week in October 2026 I asked people on X three things about AI and age. The 733 replies, from about 675 people, were each coded by stance, theme and stated age. Here is what they show.

• AI is making people over 50 more capable, not obsolete. 53% of 279 people answering my over-50 question said AI had opened doors back into work and opportunities they thought had passed them by. Another 12% said both. Only 3 people (1%) felt simply left behind.
• What AI restores is capacity. The biggest theme was people doing again what age, time or missing skills had stopped them doing. Learning, judgment and ageism followed.
• The deeper fear is not being needed. Repliers rarely feared the technology. They feared being overlooked, ending up with “nobody needing you”, and thinking less without noticing.
• Those who feel “both” are often the disrupted. People whose careers or businesses AI has shaken feel capable and threatened at the same time.
• Dependence is the real tension. Almost one in five replies to my dependence question said it is already happening. Only 19% named a clear line. The most popular rule: think first, then ask AI.
• The young are more worried than the old. National surveys show under-30s are far more pessimistic about AI and creativity than over-65s. Some older repliers worried more about their children than themselves.
• How you ask decides who answers. Posts that named an age drew warm, personal replies. A general AI post drew almost no stated ages and around four times the hostility. A personal statement earned at least three times the profile visits of a question.

The implication: the over-50 opportunity is not teaching tools. It is helping experienced people see what their judgment is worth now, who needs it, and how to use AI without handing over the thinking that made them valuable.

A question I didn’t expect to matter

Last Sunday night I posted a simple question on X. “Honest question for anyone over 50. Does AI make you feel left behind, or let back in?” I added that I had felt both in the same week. That was true.

I expected a few dozen replies. By Thursday there were 279. The post was seen more than 18,000 times. Then I posted two more. One was a statement: “I’m 69. I’m supposed to be thinking about retirement. Instead I’m thinking about what I can build next.” The other was a worry: “At what point does assistance become dependence? Where would you draw the line?”

Together the three posts drew 733 replies from about 675 people. I read every one. Then I coded them by stance, theme and stated age, with AI doing the first pass and me checking the results. What came back challenged the story we usually tell about older people and AI. It also changed how I think about what people over 50 actually need.

Why I’m in the data

I started my blog at 52, after my business fell apart. I wrote at 4:30am because it was the only quiet time I had. Nobody was waiting for a 52-year-old to reinvent himself. The internet let me do it anyway, and the blog went on to reach millions of readers.

Now I’m 69 and building again, this time with AI. So when people talk about older workers and AI, I’m not observing from a distance. I’m one of the data points. That is why I wanted to hear from others in the same boat, in their own words, rather than in a survey’s tick boxes.

Surveys tell you what people choose from a list. Replies tell you what people say when nobody gives them a list. Both matter. This piece puts the two side by side.

Where does your experience fit now?

This research is shaping what I’m building. Zyrro is an AI mentor that helps experienced people see what their experience is worth now, and who needs it.

You answer 13 questions in your own words. You get a free report on your strengths, what energises you and the themes that keep showing up in your life. The first group of testers starts soon.

Reserve your spot on the waiting list →

Finding 1: Most say AI is making them more capable

The headline is clear. Of 279 replies to the over-50 question, 53% said AI had let them back in. Another 12% said both, and 17% said neither. Only 3 people, about 1%, said they simply felt left behind.

Chart 1. Stance of 279 replies to the over-50 question.

That is not the story we usually hear. The usual story says older people are the ones AI will leave behind. The people who replied told a different story. A 57-year-old moved a 10-year-old game to a new platform. “Usually, this would have been a six-month task, now five hours.” Another reader put it more simply:

I am a woodsman that had a hatchet and now have a chainsaw.

Of course, this is not a random sample. People who reply to me on X are curious about AI. But the result still matters. Among people over 50 who are paying attention, fear is not the main feeling. Possibility is.

It also matters what “neither” meant. Most of the 17% were not afraid. They saw AI as “only a tool”, or said they had never felt out in the first place. Calm, not anxious.

Finding 2: What AI gave back was capacity

When I coded what people actually said, one theme stood far above the rest. AI had given back capacity they thought was gone. Not new ambitions so much as old ones they could finally act on.

Chart 2. Main theme of each reply, over-50 thread (one-word answers excluded).

One man approaching 60 wrote that his capacity had become more limited with age. “AI is lifting that limit.” Another reader said they could no longer work at the level of their 20s and 30s. “Now, I can get more done than I did back then.” A third could finally act on thinking they had carried for years but “lacked the skills to make it real.”

Some of the most moving replies were about identity, not output. One reader described it as “picking up a part of myself I’d put down.” A writer said AI had eased the loneliness of the craft by giving her “someone to point out your errors or inspire you to new heights.” Another wrote that AI had brought coding back into their life after eye and back strain made it too hard.

Two other themes matter for anyone over 50. The first is judgment. One reply said the tool is the part a 25-year-old can buy. “Knowing which customer exception actually counts after 1,000 tickets? Not for sale.” The second is ageism. “For the most part, AI has pushed ageism off the board,” wrote a former advertising creative director. Before AI, he said, nobody would hire you with more than 15 years of experience.

There is a money side too. One founder wrote that what once would have taken “years and millions of dollars I can now execute alone for a couple of thousand dollars.” For people with decades of knowledge and limited runway, that changes the maths of starting again.

Finding 3: Those who feel “both” are the disrupted

The 33 people who said “both” are worth a closer look. They were not confused. Many were people whose careers or businesses AI had shaken.

Chart 3. Main theme of the 33 replies that said “both”.

One wrote that AI had “caused disruption to my existing business model so am back to start-up mode at 53.” Others were learning fast but struggling to keep up. One reader summed it up in a single line:

AI can make you feel obsolete and capable again in the same afternoon.

This is the group that most needs help, and the group most likely to be missed. They are not sceptics and not cheerleaders. They are in the middle of a career change they did not choose.

Finding 4: The real fear is not being needed

My retirement post drew a different crowd. The middle stated age was 62, against 55 on the over-50 question. Most of them were in their 60s and 70s. The oldest was 80.

Chart 4. Stated ages of repliers on the two age-focused posts.

Most of these replies were celebrations. “I’m 75 and never retiring.” A 65-year-old now works as an AI enablement specialist. A 59-year-old who was called to the bar in 2023 now spends his days steeped in code and regulations, with “a 30 year work plan ahead.” A man of almost 76 is registering a new business next week. One reader’s mother, at 83, says she “hasn’t got enough time to retire anymore.”

But one reply, from a 63-year-old, named something deeper:

The script isn’t really the problem, it’s that it ends with nobody needing you.

I think he is right. The fear is not stopping work. It is stopping being needed. A 65-year-old put the same fear another way. They have the domain expertise, they wrote, “but that 65 has people not reaching out to me.” Another reader framed the opportunity: “judgment at 69 is a different asset than energy at 29.”

This lines up with wider research. In a survey of 1,505 American adults published by Elon University in September 2026, 64% agreed that by replacing humans in most tasks, AI will undermine our sense of purpose and meaning. 69% said it will weaken human bonds. Purpose, not productivity, is the deeper worry.

The tension: are we thinking less?

Not everyone was cheering. My third post asked where people would draw the line between assistance and dependence. Its 282 replies were the most divided of the three.

Chart 5. Replies to “Where would you draw the line?”

Almost one in five said dependence is already happening. One reader watched colleagues “who far outrank me in experience and education” struggle “to think even simple concepts without asking AI.” Another admitted: “Forgetting how I once did things.” Several used the GPS comparison. Use it or lose it. One reader put it bluntly: “AI is doing to thinking what social media did to attention spans.” A 56-year-old engineer said we crossed that line decades ago, when schools stopped asking students to show their working.

Others said the opposite. A retired reader wrote that he had “used more math in the last 12 months than I did in the preceding 30 years of career.” Another said he had never read so much in his life. Same tool. Opposite result.

One reply reframed the problem. Dependence, he argued, is not driven by laziness but by “impatience and anxiety”. The world moves too fast, so people reach for the fastest answer. Another offered a simple test for when you have crossed the line: when someone asks you a question and you “can’t answer it without having the urge to use AI.”

The same Elon survey captures both sides. Among people who use AI chatbots, 51% said they feel lazy or take shortcuts at least some of the time, and 39% said they are growing too dependent. Yet 52% said AI had improved their motivation to learn new skills, against 5% who said it made it worse.

The research suggests the difference is how you use it. In a study by Microsoft and Carnegie Mellon of 319 knowledge workers, more trust in AI went with less critical thinking, while more self-confidence went with more. In a small MIT Media Lab study (54 people, not yet peer reviewed), most people who wrote essays with ChatGPT could not quote their own work afterwards. The group that wrote first and used AI second stayed the most engaged.

The most-liked reply in the thread, with 21 likes, said the same thing in one line. “Formulate your own ideas first, then use AI to challenge, refine and expand.” Other readers had their own versions. One asks AI to find sources she then reads herself. Another never uses AI on a topic they don’t already know, because they couldn’t spot its mistakes. A writer never uses AI to write, because “what sells is my style.” Someone whose agency runs 90% on AI agents said the line is “taste and final accountability.”

One reader made the point with humour. He asked AI where to draw the line. “It recommended independent thinking, then offered to do it for $20 a month.”

The young are more worried than the old

Here is a twist the headlines miss. Older people are not the most worried about AI. The young are.

Chart 6. Pew Research Center, US adults, 2025.

The Pew Research Center found that 61% of Americans under 30 expect AI to make people worse at thinking creatively. Among those 65 and older, it was about four in ten. Under-30s were also more likely to expect AI to damage relationships.

My replies showed the same pattern from the other side. Some of the deepest worry was not for themselves. “What actually keeps me up is my adult kids,” one parent wrote. “They’re trying to get a first job competing against agents.” Another reader turned an old joke around. Parents used to ask their kids for tech help. Now, he said, “they now come to me when they need help using it.”

There is resentment too, and it deserves honesty. One younger reply said: “Because you won’t retire, an entire generation is struggling.” That is a real tension. If experienced people now have more leverage, the question of how they make room for the next generation becomes more urgent, not less.

A surprise: how you ask decides who answers

The three posts drew three different crowds. The two posts about age drew people who shared their age and their stories. The general post about AI dependence drew almost no stated ages and around four times the hostility. One person replied with two words: “Thanks Gramps.”

Chart 7. Share of replies stating an age, and share dismissive or hostile.

The posts also performed differently. The personal statement about retirement earned 19.6 likes and 15.6 profile visits per 1,000 views. The questions earned more replies but far fewer visits. A question starts a conversation. A personal story makes people want to know who you are.

Chart 8. Likes, profile visits and replies per 1,000 views.

There is a lesson here beyond social media. If you are over 50 and want to be found, lead with your story and your age. Hiding it attracts the wrong crowd. Owning it attracts your people.

What the cheerleaders miss

Not every reply was a victory lap, and the doubts deserve space. One man has been retired for 17 years and is “happily doing basically nothing.” Another wants naps, a cello and the beach. They are not wrong. As one reader put it, retirement “was designed around a world where capability faded with age and tools stayed the same.” Retirement is not the problem. Being told when to do it is.

Others warned about the cost of too much possibility. One reader took a month off after pushing himself to “near mental exhaustion.” Another warned: “Watch the burnout, this fun can quickly take over.” A third described building as “a new drug.” He said he was addicted to “building shit nobody wants because I can now.” AI makes building easy. It does not make finding the people who need it any easier.

And there are practical barriers. A 70-year-old wrote that coming out of retirement would cost him more than $80,000. For many people, the choice to keep working is shaped by tax, pensions and health, not just by mindset.

Six kinds of people in the replies

Reading 733 replies, six groups kept appearing. They are not neat boxes, and many people fit more than one. But they help explain why the same technology feels so different to different people.

1. The capacity restorers

The largest group. Age, energy or missing skills had limited them, and AI has lifted the limit. They want to do more of what they already know.

2. The rediscoverers

People picking up an old passion or an old self: music, coding, writing, a game they built a decade ago. For them AI is less a tool than a door back to who they were.

3. The solo founders

People building a business alone that once needed a team and capital. They are energised, but some are building without a clear customer.

4. The disrupted reinventors

People whose careers or businesses AI has shaken. They often answered “both”. They need direction more than tools.

5. The overlooked experts

People with deep knowledge who feel invisible. AI gives them power, but not yet an audience or a market. Their fear is not being needed.

6. The content retirees and sceptics

People who are happy as they are, or who see AI as just a tool. They are a useful check on the hype, and they deserve respect, not persuasion.

Why it matters

Put these findings together and a clear picture emerges. Most people over 50 in this sample are not afraid of AI. They are afraid of being overlooked, of no longer being needed, and of thinking less without noticing.

That is a very different problem from “learn to use the tools.” The tools are the easy part. The hard part is knowing what your experience is worth now, who needs it, and how to use AI without handing over the thinking that made you valuable in the first place.

For employers, the lesson is that experience plus AI may be the most underpriced combination in the workforce. For creators and builders, the over-50 audience is not a group to be rescued. It is a group to be served, with different needs in each of the six groups above. For the rest of us, it is a reminder that a long life of learning does not have to end with a script someone else wrote.

What to do with this

1. Think first, then ask. Write your own view before you open the AI. Use AI to challenge it, not replace it.
2. Name your judgment. Write down the decisions you can make that a 25-year-old with the same tools cannot. That list is your edge.
3. Lead with your story. Your age and experience attract the right people. Hiding them attracts the wrong ones.
4. Build for someone. Before you build, name the person who needs it. Possibility without a customer becomes a hobby.
5. Find who needs you. Make a list of people your experience could help this year: a younger colleague, a small business, your own children.
6. Keep one skill manual. Choose one thing you will always do yourself, whether it is writing, maths or reading a map. Treat it like exercise.
7. Protect your energy. The excitement is real. So is burnout. Decide your stopping time before you start.

How this research was done

This is a small, honest study, not a poll. The replies came from three of my posts on X between 4 and 8 October 2026: the over-50 question (279 replies from 262 people), the retirement statement (172 from 165) and the dependence question (282 from 270). I collected every visible reply on 8 and 9 October.

Each reply was given one stance and one main theme. AI did the first pass and I checked the results by hand. Ages were counted only where people stated their own age, so the age figures cover 36 and 31 people respectively. Engagement figures come from my X analytics export.

The limits matter. Repliers chose to respond, so the sample leans towards people curious about AI and towards people who already follow me. About six in ten of the people who like my posts are 55 or older and about six in ten are in the US. About eight in ten are men. A survey of Facebook users or retirees offline might look very different. Quotes are from public replies, and I have left out names.

One question to leave you with

A reader who had just turned 51 told me that time usually speeds up as you get older. With AI, he wrote, “for me it feels like it is slowing down.” More gets done in a day, so the days feel bigger.

I feel that too. At 69 I have more ideas than time, and for the first time in years the tools are not what is holding me back. The real question is the one my 63-year-old reader raised.

If AI has given you back some capacity, who will you use it for?

Find out what your experience is worth now

The people in this research didn’t ask for more AI tools. They asked a harder question: what is my experience worth now, and who needs it?

That’s the question Zyrro is built to help you answer.

1. Answer 13 questions in your own words. Zyrro asks before it suggests anything.
2. Get your free report: your strengths, what energises and drains you, and the patterns that keep showing up.
3. Explore four possible paths for your next chapter, with a seven-day plan to test one in the real world.

It’s designed so you think first and need it less over time. The first group of testers starts soon.

Join the Zyrro waiting list at zyrro.ai →

The post What 733 People Over 50 Told Me About AI (Original Research) appeared first on jeffbullas.com.

* This article was originally published here

Start making $100+ per day this week with the best dfy system - Subscribe here!

The One-Person AI Company’s Biggest Problem Is Being Found

In space, no one can hear you scream. In the AI economy, no one can hear you launch.

There is a new dream moving through the technology world: one person, one laptop and one idea.

No large development team. No expensive agency. No office full of engineers. Agentic AI can now help a founder research a market, shape a product, write code, design the interface, test it, build the website, create content, handle support and automate parts of the operation.

A company that once needed twenty people can increasingly be attempted by one motivated founder with a collection of AI agents beside them. That is extraordinary.

But there is a problem hiding behind the excitement. You can build the product, launch the website and switch on the payment system. Then comes the silence.

The world does not automatically arrive. You have created something in a universe already crowded with millions of other things being created at the same time. You are shouting into the vacuum.

Why this matters

If you are building the company and its easy then millions more are doing the same thing. The AI era is both creating opportunity while reducing its visibility when you launch. 

This may become the defining paradox of the one-person AI company. AI is removing the bottleneck of creation while increasing the bottleneck of discovery.

For most of the software era, building was hard. Distribution was hard too, but if you could raise the money, assemble the programmers and ship something useful, you had already crossed a major barrier.

Now that barrier is collapsing. AI is making production abundant, so scarcity moves somewhere else: attention, trust, reputation and distribution.

The one-person AI company revolution is not complete when one person can build a company. It is complete when one person can reliably find the customers who need what they built.

The old internet rewarded publishing

For roughly two decades, the web operated under an informal bargain. You created useful information, search engines indexed it, social networks distributed it and people clicked.

Some became readers. Some became subscribers. A few became customers. That model helped blogs become media companies, creators become brands and tiny software businesses become global products.

I experienced that era personally. Start a blog. Publish relentlessly. Earn links. Build an email list. Grow an audience. Let search and social platforms carry your work farther than you could ever carry it yourself.

It was never easy, but there was a visible path. That path is becoming less predictable.

Pew Research Center analysed 68,879 Google searches made by 900 U.S. adults. When no AI summary appeared, 15% of search visits led to a click on a traditional result. When an AI summary appeared, that fell to 8%. Only 1% clicked a source link inside the AI summary. Read the Pew study.

Figure 1. AI summaries can satisfy the query before the user reaches the original source.

That number matters because it changes the economics of being useful.

Your article can be researched, your insight can be extracted and your answer can be summarised. The user can receive the value without ever reaching the original source.

Being used as a source is no longer the same thing as being discovered as a brand. For a one-person company, that difference can decide whether valuable work produces customers or simply disappears into someone else’s answer.

The invisible villain

There is a villain in this story, but it is not a single company. It is a system built around convenience, speed and keeping the user inside the interface.

The system wants to reduce friction. It wants to answer faster. It wants people to keep scrolling, searching or asking questions without leaving.

For the user, that can be wonderful. For the unknown founder, it can create a strange outcome: your knowledge becomes useful while your name becomes optional.

Nobody designed the entire effect from one room. Search engines, social platforms, recommendation algorithms and AI assistants are all responding to powerful incentives.

But the result is the same. 

You can be valuable and invisible at exactly the same time.

The Reuters Institute reported that Google organic search referrals to more than 2,500 news sites fell 33% globally between November 2024 and November 2025, and 38% in the United States. It also notes that the figures do not prove AI Overviews caused the entire decline; algorithm changes and changing search behaviour matter too. See the 2026 Digital News Report.

Figure 2. The open web’s old referral engine is under pressure.

The exact numbers will vary by sector. A news publisher is not a SaaS company, a local accountant is not a creator and a niche B2B startup is not an ecommerce store.

Still, the direction is difficult to ignore. Distribution that once felt almost automatic now needs to be designed with much more intention.

For the one-person AI founder, relying on one platform or one algorithm is becoming a fragile strategy.

AI can consume more than it returns

Cloudflare looked at crawler requests compared with attributable human referrals in June 2025. Its estimate was roughly 14 crawler requests per referral for Google, about 1,700 for OpenAI and about 73,000 for Anthropic. Cloudflare cautions that referrals from native AI apps may be undercounted because they may not pass a standard referrer header. Read Cloudflare’s analysis.

Figure 3. Crawling and referral are no longer the same exchange.

The caveat matters. 

Crawling is not the same thing as citing, training is not the same thing as search retrieval, and those ratios are not a moral scorecard.

But they do reveal a structural change. The web’s old bargain “Crawl my content and send me humans” is no longer guaranteed.

For the one-person AI company, this means that “publish more” is not enough. 

You need to create something the machine cannot completely absorb and deliver on your behalf.

So what becomes scarce?

Information is no longer scarce. It is exploding.

AI can generate explanations, lists, summaries, plans, landing pages, newsletters and social posts in seconds. If your entire value proposition is information, you are building on ground that is getting cheaper every month.

What becomes more valuable is what cannot be neatly compressed into an answer. That includes original evidence, lived experience, judgment, trust, relationships, implementation and transformation.

• Original evidence: data you collected, tests you ran, outcomes you measured.
• Lived experience: what happened when you actually tried the thing.
• Judgment: what matters, what does not, and what you would do next.
• Trust: the belief that you will deliver what you promise.
• Relationships: access to people who know you, reply to you and recommend you.
• Implementation: helping someone move from knowing to doing.
• Transformation: a measurable change in the customer’s life or business.

This is the part of the one-person AI company story that interests me most.

AI can help one person create at the scale of a team, but the founder still has to become a signal inside a world filled with synthetic noise.

That is not mainly a technology problem. It is a human problem: how to be noticed, believed, remembered and recommended.

The answer is not to crack the algorithm

Founders often ask the wrong question. 

• How do I crack LinkedIn? 
• How do I game Google? 
• What is the perfect posting schedule on X? 
• What is the newest SEO trick?

The problem with that strategy is simple. You do not own the algorithm. You are renting distribution from a landlord who can change the lease overnight.

The better question is more durable: how do I build a distribution system that becomes more valuable even when algorithms change?

Borrow attention. Own the relationship. Earn advocacy.

That suggests a three-part strategy.

1. First, use platforms you do not own for discovery. Search, X, LinkedIn, YouTube, Reddit, podcasts, AI assistants and other people’s audiences can all introduce you to the right people.
2. Second, create a direct relationship. Email, a product account, a community or a recurring service gives that person a way to find you again without asking an algorithm for permission.
3. Third, deliver enough value that customers become distribution. They tell colleagues, share an output, provide a review or introduce you to a partner.

That is when distribution starts to compound.

Borrow trust before you build an audience

The other mistake is assuming every one-person founder needs to become an influencer. They do not.

Imagine you have built a remarkable AI tool for independent financial advisers. You could spend two years trying to build 100,000 followers.

Or you could find ten consultants, newsletter writers, associations, podcasts or software companies that already have the trust of financial advisers.

One trusted introduction to 500 people with the exact problem you solve may be worth more than 500,000 random impressions. This is borrowed distribution, but more importantly it is borrowed trust.

For an unknown one-person company, trust is often the missing bridge between being seen and being tried.

Build something AI cannot replace with a paragraph

There is another test I would apply to every new AI company: can the customer get most of your value by asking an AI assistant a good question?

If the answer is yes, the moat is thin. A generic career guide can be summarised, a list of marketing ideas can be generated and a startup plan can be produced.

A system that understands the customer, watches what they do, pushes them to act, measures what changes and adapts to the result is different. The value moves from information to experience.

It moves from advice to implementation, and from possibility to progress. That is much harder to steal with a summary.

There is also an unexpected opportunity

Adobe Digital Insights found something that complicates the doom story. In U.S. retail data, AI-referred traffic converted 38% worse than non-AI traffic in March 2025. By March 2026 it converted 42% better. Adobe also reported higher engagement among AI-referred visitors. See Adobe’s 2026 analysis.

Figure 4. AI can reduce some clicks while increasing the intent of the clicks that remain.

Retail is not SaaS, and one study should not be treated as a universal law. But the implication is fascinating.

AI may not simply become a wall between you and the customer. It may become a new discovery layer.

Someone asks, “What is the best tool for an experienced executive who wants to decide what to do next?” Or, “What software helps a solo consultant automate client reporting?”

If an AI assistant knows your product, understands what it does and can find independent evidence that it works, that recommendation may become extremely valuable.

The goal is no longer merely to rank in Google. The new goal may be to become recommendable by machines and trusted by humans.

The secret sauce: a one-person distribution engine

So how does the invisible one-person AI company become visible?

Not with one growth hack, a magic posting schedule or another automated content machine. It becomes visible by building a loop in which discovery, trust, experience, evidence and advocacy reinforce each other.

Figure 5. The six-step one-person AI company distribution engine.

That is the shift.

The one-person AI company should not think like a miniature corporation. It should think like a network, with the founder in the middle, AI agents around them and customers, partners and communities around the outside.

Every useful interaction creates another signal. Every successful customer creates another piece of evidence. Every trusted partner creates another doorway.

Distribution stops being an event that happens after launch. It becomes part of the product architecture.

The second revolution of the one-person company

The first revolution is already underway: one person can increasingly do the work of many. But that is only half the story.

The second revolution will be about whether one person can also build a distribution system that once required a marketing department, a sales team, a PR firm and a media budget.

Agentic AI will help here too. Agents can monitor markets, identify prospects, research partners, repurpose original research, personalise onboarding, analyse conversion and reveal where customers drop away.

But automation cannot manufacture trust. It cannot fake a genuine customer outcome, create a reputation that has not been earned or save a product nobody actually needs.

Those remain human constraints. The future may belong not to the founder who automates everything, but to the founder who knows what should never be automated.

Creation is becoming abundant. Discovery is becoming scarce. Trust may become the ultimate distribution advantage.

The one-person AI company is real, but the fantasy is believing that building it is enough.

The hard part is no longer merely making the thing. The hard part is becoming visible to the right people, being credible when they arrive and delivering an outcome strong enough that they tell someone else.

That is how an invisible company becomes a business.

Because in space, no one can hear you scream. And in the emerging AI economy, no one can hear you launch unless you give the right people a reason to listen.

Research sources

• Pew Research Center: Google users are less likely to click on links when an AI summary appears
• Reuters Institute: Digital News Report 2026
• Cloudflare: Control content use for AI training
• Adobe: AI traffic grows but retail sites lag in AI search visibility

The post The One-Person AI Company’s Biggest Problem Is Being Found appeared first on jeffbullas.com.

* This article was originally published here

Start making $100+ per day this week with the best dfy system - Subscribe here!

AI Didn’t Kill Authorship. It Changed What Authorship Means.

About ten years ago I was wandering through the Musée Rodin in Paris. I am not an artist. I went in with the naive assumption most of us carry into a gallery: Rodin made the Rodins.

Then I discovered the studio behind the name.

Rodin was deeply involved in the conception and modelling of his sculpture, but the physical realization of major works often depended on a network of assistants and specialist craftspeople. The Musée Rodin explains that he employed skilled practitioners to enlarge and reduce models, marble carvers to translate plaster into stone, foundries to cast bronze, and assistants such as Camille Claudel who worked on difficult details. The hand that struck every blow of the chisel was not always Rodin’s. Yet we still say: “a Rodin.”

That discovery stayed with me because it disturbed a simple idea of authorship I had never examined.

Then, a few days ago, I read a New York Times profile titled “The Incredible Art of Being Jeff Koons”. The same question returned, but in a form that feels remarkably relevant to the age of AI.

At one point Koons had more than 100 people working in his studio, including teams hand-copying old masters. Today the operation is smaller, but the system remains exacting: digital models, scans, standardized lighting, paint-by-numbers codes, external foundries and specialist fabricators. Koons told the Times that his systems are intended to ensure that every gesture, colour and shape is the way he wants it.

Who, then, made the art?

The easy answer is: his team. The more interesting answer is: Koons authored it.

That distinction may help us understand one of the most uncomfortable questions facing writers now: if an AI helps research, structure, edit and even produce sentences, who is the author?

The false choice: purity or abdication

The argument around AI and writing has hardened into two camps.

At one extreme is abstinence. A “real” writer should write every word. AI may be allowed to fix spelling or perhaps find a source, but the moment it generates prose the work becomes suspect.

At the other extreme is abdication. Give the machine a topic, ask it to research, outline, draft and polish 1,500 words, glance over the result, then publish it under your name.

The first confuses authorship with keystrokes. The second confuses supervision with authorship.

The scale of the temptation is easy to understand. In a randomized experiment published in Science, Shakked Noy and Whitney Zhang gave 453 college-educated professionals realistic writing tasks. Access to ChatGPT reduced average completion time by 40 percent and increased assessed quality by 18 percent. That is not a marginal productivity improvement. It is a reason every writer, marketer, consultant and knowledge worker is being forced to rethink how work gets made.

Figure 1. ChatGPT’s measured effect on professional writing tasks. Source: Noy & Zhang, Science (2023).

But speed is not the same as authorship.

From maker to architect

Humans have been moving up the abstraction stack for centuries.

We moved from muscle to machines; from doing every task to operating the machine; from operating the machine to designing the system. The builder becomes the architect. The craftsperson becomes the creative director. The founder who once needed a department can increasingly orchestrate a network of software and AI agents.

AI accelerates that move because it can automate not only physical labour but parts of cognitive labour.

For writers, the ladder might look like this: sentence maker → editor → director → architect of meaning.

There is enormous leverage in that progression. One person can explore more research, test more counterarguments, surface more connections and move from idea to published work faster than ever before.

But there is a trap in the metaphor of “moving up.” Higher leverage does not automatically mean higher mastery.

Rodin could delegate because Rodin knew sculpture. He had modelled clay, studied bodies, understood proportion and developed taste through practice. That accumulated craft gave him the capacity to look at work produced by someone else and say: no, that is wrong.

The same is true of a great editor, architect, chef or conductor. Direction is strongest when it rests on an internal model built by doing the work.

AI now allows us to leap up the leverage ladder before we have climbed the mastery ladder. That creates a new figure: the fragile director. They can generate impressive-looking work at extraordinary speed but may not possess the craft required to tell whether it is actually good.

Making is also thinking

This is why I resist the idea that the ideal future is one in which humans stop doing and spend all their time directing machines.

Part of what I love about art is technical excellence: seeing what a human hand has learned to do with marble, pigment, a brush, wood, sound or light. Sometimes the execution is not merely a delivery mechanism for the idea. The execution is part of the idea.

The same applies to writing.

A sentence can contain information, but it can also contain rhythm, restraint, surprise, personality and years of practice. We admire Orwell or Didion not merely because of what they thought, but because of the precision with which they learned to express thought.

The danger is that AI can give us the appearance of mastery without the apprenticeship that produces judgment.

A 2026 review in Trends in Cognitive Sciences describes this broader problem as cognitive offloading. Offloading work to AI can impede skill acquisition or contribute to skill decay, although the authors stress that the outcome depends on how the technology is used. That qualifier matters. The issue is not AI itself. It is what we choose to stop practising.

Automate friction that wastes life. Preserve friction that develops you.

Transcribing a two-hour interview by hand is mostly friction. Searching 100 documents for a quote can be friction. Reformatting citations is friction.

But wrestling with an argument can be formative friction. Finding the sentence that says exactly what you mean can be formative friction. Writing an introduction before asking AI for one can expose what you actually think.

The AI creativity paradox

The research increasingly suggests that AI creates a trade-off rather than a simple win or loss.

In a Science Advances experiment, Anil Doshi and Oliver Hauser gave writers access to generative-AI story ideas. Writers who could request up to five AI ideas produced stories rated 8.1 percent more novel and 9 percent more useful than the human-only group. The biggest gains went to writers who started with lower creativity scores.

But the stories also became more alike. AI-assisted stories moved closer to the average story in their condition. With access to one AI idea, the increase in similarity represented 10.7 percent of the similarity-score range found in the human-only group.

Figure 2. One AI-generated idea improved judged creativity while also increasing similarity between stories. Source: Doshi & Hauser, Science Advances (2024).

That paradox has since become harder to dismiss. A 2025 study of 2,200 college admissions essays found that human-written essays added new semantic diversity roughly two to eight times faster than base GPT-4 essays as the number of essays increased. And a 2026 meta-analysis spanning 19 studies and 61 effect sizes found a small but statistically significant homogenisation effect in human-AI co-creation.

AI can make each of us better while making all of us more similar.

That is where “AI slop” enters the story.

In 2025, Merriam-Webster named “slop” its word of the year and defined it as low-quality digital content produced, usually in quantity, with AI. The phrase is useful, but I think slop is a symptom rather than the disease.

The deeper problem is abdication.

AI slop appears when we outsource not only production but curiosity, experience, point of view, taste and judgment. The machine supplies the topic, the structure, the examples, the language and sometimes even the conclusion. The human becomes a publishing endpoint.

The result can be grammatically clean and intellectually empty.

A new definition of authorship

This is why Rodin and Koons matter to the AI writing debate.

They show that authorship has never required the author to perform every physical act of production. Art has a long history of workshops, apprentices, assistants, foundries and specialist fabricators. What matters is the nature of the contribution and the degree of creative control.

Even copyright law is moving toward this distinction. In its 2025 report on AI and copyrightability, the U.S. Copyright Office concluded that using AI as an assistive tool does not prevent copyright protection. Human-created selection, arrangement or modification can qualify. But simply providing prompts is not, by itself, enough to establish authorship of AI output.

That is a legal standard, not a complete philosophy of writing. But the direction is useful.

Authorship is not “I touched every word.”

It is closer to five responsibilities: Origin — why does this work exist? Intent — what am I trying to say? Direction — what should be researched, included, excluded or challenged? Judgment — is this true, interesting, beautiful, useful and mine? Responsibility — am I prepared to put my name behind it and defend it?

None of those requires typing every sentence. All of them require being present.

The two ladders of AI-assisted creation

I now think creators need to climb two ladders at once.

1. The first is the leverage ladder: maker → operator → director → architect. 
2. The second is the mastery ladder: novice → apprentice → craftsperson → master.

AI can rocket us up the first ladder. It cannot automatically carry us up the second.

Figure 3. The strongest AI-age creator combines high leverage with high mastery. Framework: Jeff Bullas.

The dangerous position is high leverage and low mastery: the fragile director.

The exciting position is high leverage and high mastery: the master-director. That is Rodin with a studio. It is the architect who understands construction. It is the editor who has written thousands of pages. And it may be the strongest model for the AI-age writer.

The boundary I am trying to draw

I am still working this out in my own writing.

Sometimes I have abdicated too much. I have supplied a topic or headline and let AI run too far. Other times I have written the opening, supplied the lived experience, directed the research, challenged the argument, rejected language, moved sections and edited heavily. Increasingly, I think of the second approach as authorship rather than purity.

The percentage of AI-generated words is a poor test.

A better test is whether the work would exist in substantially the same form without the human behind it.

Did the piece begin with something I noticed, experienced or genuinely wanted to understand? Did I decide what question mattered? Did I challenge the evidence? Did I choose what belonged and what did not? Could I explain and defend the argument without opening the AI chat? Would another person giving the model the same headline have produced essentially the same article?

And one more question may matter even more: Am I still practising the craft that allows me to judge the machine?

I don’t want AI to free me from writing. I want it to free me from unnecessary labour so I can spend more time on observation, thought, story, craft and judgment.

The future should not be a civilisation of people who have forgotten how to make things but have become excellent at requesting them.

Nor should we romanticise unnecessary labour merely because humans once had to perform it.

The better destination is the master-builder: hands capable of making, a mind capable of designing, judgment capable of directing, and technology capable of multiplying all three.

Use AI to expand thought and amplify expression. Do not let it decide what you mean.

Or even more simply:

Delegation can expand authorship. Abdication abandons it.

The deeper question: what will you do with all this leverage?

AI can help us write faster.

It can help us research more deeply, explore more options and produce at a scale that was impossible a few years ago.

But that creates a new problem.

The more capability we gain, the more important it becomes to know:

What do I actually want to create?

What is worth my attention?

What deserves my time, energy and commitment?

The danger is not only that AI starts writing for us.

It is that we become surrounded by so many possibilities that we lose sight of our own direction.

That is part of why I’m building Zyrro.

Zyrro is designed to help you understand the patterns behind who you are, what energizes you, what matters to you and which paths may be worth exploring next.

Not to hand you a fixed answer.

Not to tell you what your purpose is.

But to help you make better choices in a world where AI can generate almost infinite options.

Because the real opportunity of AI is not simply to produce more.

It is to give us more leverage to become more intentional about what we choose to make, pursue and become.

AI can amplify your capabilities. Zyrro is being built to help you decide where to point them.

If that sounds useful, join the Zyrro waitlist and follow the journey as we launch.

Research & source links

• Musée Rodin — Multiples, fragments, assemblages
• Musée Rodin — Marble carving
• Musée Rodin — Camille Claudel
• The New York Times — The Incredible Art of Being Jeff Koons
• Noy & Zhang (2023), Science — productivity effects of generative AI
• Doshi & Hauser (2024), Science Advances — creativity and collective diversity
• Moon, Green & Kushlev (2025) — homogenizing effect of LLMs on creative diversity
• de Rooij & Biskjaer (2026) — meta-analysis of homogenisation in human-AI co-creation
• Cash et al. (2026), Trends in Cognitive Sciences — cognitive offloading and skill decay
• U.S. Copyright Office — Copyright and Artificial Intelligence, Part 2
• AP — Merriam-Webster’s 2025 word of the year: slop

The post AI Didn’t Kill Authorship. It Changed What Authorship Means. appeared first on jeffbullas.com.

* This article was originally published here

Start making $100+ per day this week with the best dfy system - Subscribe here!