For years, the internet worked in a relatively simple way.
You had a question. You opened a search engine, typed a few keywords, looked through a list of results, and clicked the websites that seemed useful. You compared sources, read explanations, followed links, and gradually built your own understanding.
Today, that experience is changing.
Instead of giving you a list of websites, AI-powered search experiences can generate summaries, explain concepts, compare options, and answer questions directly. You can ask a follow-up question without opening another page. You can request a summary of a complicated subject and receive one in seconds.
It feels like progress, and in many ways, it is.
Finding information is becoming faster and more convenient. But as the internet becomes better at delivering answers, I think we should ask an important question:
If we no longer need to visit websites to get answers, what happens to the websites that make those answers possible?
This is not just a debate about search engines or artificial intelligence. It is about how knowledge is created, how independent publishers survive, how developers discover solutions, and how the next generation of the internet will work.
1. Search Was More Than Finding an Answer
Traditional search engines did not simply provide information. They provided a path to information.
When you searched for a programming error, you might find an official documentation page, a GitHub issue, a Stack Overflow discussion, a personal blog, or a tutorial written by someone who had encountered the same problem.
Each source offered a different perspective.
Official documentation explained the intended behavior. Community discussions revealed confusing edge cases. Personal blogs described practical mistakes. GitHub issues sometimes exposed bugs that were not documented anywhere else.
You had to navigate these sources, compare explanations, and decide which information applied to your situation.
That process could be frustrating, but it also exposed you to information you did not initially know you needed.
An AI-generated answer changes this experience. It can combine information into a single explanation, removing much of the effort required to find and compare sources.
For straightforward questions, this is genuinely useful.
But the process of searching, comparing, and reading also served another purpose: it connected readers with the people and communities that created the information.
When that connection disappears, the consequences extend beyond convenience.
2. AI Answers Are Changing How People Visit Websites
This shift is already visible in research on online browsing behavior.
A Pew Research Center study published in July 2025 analyzed 68,879 Google searches conducted by 900 US adults during March 2025. The researchers found that users who encountered an AI summary clicked a traditional search result in 8% of visits, compared with 15% of visits without an AI summary. Links within AI summaries were clicked in just 1% of visits where summaries appeared.
Pew Research Center
+1
Research context: These figures describe the observed browsing behavior of the study participants during the study period. They are not universal click-through rates for every search engine, country, or type of query.
The findings illustrate an important change in user behavior.
When an answer appears directly on the results page, some users have less reason to visit another website.
Imagine searching for a simple technical question, such as how to convert a string to an integer in Python. An AI summary might provide a working example, explain the function, and mention a common error.
For a quick task, that could be everything you need.
Now consider a more complicated question: how should you design authentication for a production application?
A short answer might explain the basic concepts, but you may still need official documentation, implementation examples, security guidance, and discussions of edge cases.
The challenge is that AI search can present both kinds of answers in a similarly convenient format. Users may receive a concise explanation without immediately knowing what important context is missing.
The convenience is real. So is the possibility that useful sources receive fewer visits.
3. What Happens to Independent Publishers?
Websites require time, effort, and often money to maintain.
Independent developers write tutorials. Researchers publish explanations. Journalists investigate stories. Educators create learning materials. Small businesses maintain documentation to help customers solve problems.
Many of these creators depend on some combination of advertising revenue, subscriptions, donations, product sales, consulting, and referrals.
Search traffic can help readers discover their work.
If an AI system uses information from across the web to generate an answer, but fewer users visit the original websites, the relationship between information creation and its economic rewards becomes more complicated.
A creator might invest hours testing a solution, documenting an edge case, and publishing a detailed tutorial. An AI-generated summary could communicate the central idea in a few sentences, satisfying a user's immediate need without generating a visit to the original article.
This does not mean every AI answer replaces a website visit. Some people will still click through to read detailed explanations, verify sources, or explore related material. Search behavior also varies by topic and intent.
However, the possibility of reduced referral traffic raises a difficult question.
If creators receive less traffic from the systems that use their work, how will they continue producing high-quality information?
This is not only a concern for publishers trying to protect their income. It affects the availability of future knowledge.
A healthy internet needs people who are willing and able to investigate problems, test solutions, maintain documentation, and share what they learn.
If those activities become harder to sustain financially, the long-term effects could reach well beyond individual websites.
4. The Risk of Losing Original Experience
There is a difference between knowing an answer and understanding how someone arrived at it.
Consider a developer who publishes a tutorial about debugging a database connection problem.
The article might explain the error message, but its real value could be in the details: the operating system, the database version, the failed configuration, the misleading error, and the exact change that fixed the problem.
These details often come from experience rather than simply knowing the correct command.
An AI system can summarize the explanation, but a short summary may leave out the conditions under which the solution works. If the reader receives only the summary, they might miss the details needed to apply it safely.
This becomes particularly important in software development.
A code snippet that works in a demonstration may fail in production because of concurrency, permissions, input validation, performance constraints, or differences between software versions.
Original articles often preserve the reasoning behind a solution, including failed attempts and lessons learned.
For example, a developer writing about a URL shortener might explain why a database constraint was necessary to prevent duplicate codes, how invalid URLs were handled, and what happened when two requests attempted to create the same identifier.
Those experiences provide context that is difficult to capture in a generic answer.
AI summaries can help people access information more quickly. But we should not confuse a shorter explanation with a complete understanding.
5. The Internet Could Become More Convenient but Less Diverse
Another potential loss is diversity of perspective.
When you visit several websites, you encounter different writing styles, opinions, priorities, and ways of explaining the same problem.
One developer may prefer a minimal implementation. Another may emphasize security. A third may explain the historical reasons behind a design decision.
Reading these different perspectives helps you recognize that technical problems rarely have only one reasonable solution.
AI-generated answers can bring multiple perspectives together, but they can also present a single synthesized explanation that hides the disagreements or uncertainty behind it.
This creates a subtle risk.
Readers may become accustomed to receiving one polished answer instead of exploring the range of ideas available across the web.
That does not mean AI answers are inherently less diverse. Their quality depends on the sources they use, how they synthesize information, and whether they accurately communicate disagreement and uncertainty.
The concern is what happens when the summarized answer becomes the user's entire information experience.
A search engine that directs you to several sources invites exploration. An answer engine can make exploration feel unnecessary.
For routine questions, that may be a reasonable trade-off. For subjects involving security, public policy, scientific uncertainty, or important personal decisions, the ability to examine original evidence remains valuable.
6. What Developers Could Lose
For developers, the changing relationship between search and websites deserves particular attention.
The open web has traditionally provided a large, distributed knowledge base. Developers can search error messages, inspect code examples, read official documentation, and learn from discussions about unusual bugs.
Much of this knowledge exists because people chose to publish their experiences.
A developer who encounters a strange framework issue might discover a small personal blog describing the exact problem. That blog may have limited traffic, but it can be extremely useful to someone facing the same issue.
If discovery increasingly happens through generated answers, smaller sources could become harder to find, especially when their content is summarized without a strong reason for users to visit the original page.
https://goodoff.co/
There is also a potential feedback problem.
Developers publish solutions and documentation.
AI systems use publicly accessible information to help answer questions.
Users receive answers without necessarily visiting the original sources.
Some creators receive less traffic or fewer opportunities to earn revenue.
Fewer creators may have the resources or motivation to maintain detailed public resources.
This is a possible risk, not an inevitable outcome. AI tools can also direct users to useful documentation, introduce readers to unfamiliar projects, and help more people discover technical knowledge.
The question is whether those benefits will be sufficient to sustain the communities that produce the information.
For developers who rely on the open web, preserving access to original documentation, code repositories, issue discussions, and tested examples remains important.
7. Can AI Search and the Open Web Coexist?
I do not think the solution is to reject AI-powered search.
There are clear benefits to generating direct answers. People can find information faster, ask follow-up questions naturally, and get explanations adapted to their level of understanding.
AI can also help users discover relevant concepts they might not have known how to search for.
The challenge is designing these experiences so that convenience does not come at the expense of the wider information ecosystem.
There are several practical ways to move in that direction.
Make Sources Visible and Useful
An AI-generated answer should make it easy to identify and visit the sources behind important claims.
Source links should lead to the relevant material, not merely a generic homepage. Readers should be able to distinguish between information supported by original research, official documentation, and secondary commentary.
Citations are most valuable when they help users investigate a claim rather than simply decorate an answer.
Give Original Work a Reason to Be Visited
Websites can provide value that is difficult to reproduce in a short summary.
For technical publishers, that could mean runnable examples, downloadable projects, interactive demonstrations, benchmark results, detailed experiments, and explanations of failure cases.
For researchers and journalists, it might mean access to underlying data, methodology, interviews, original documents, and detailed reporting.
The objective should not be to make information unnecessarily difficult to access. It should be to make the original resource useful beyond the basic answer.
Measure More Than Convenience
Search systems are often evaluated by how effectively they satisfy a user's information need. That is important, but the broader ecosystem matters too.
It is also worth asking whether the system directs users toward reliable sources, preserves attribution, supports independent publishers, and encourages further investigation when a question requires more depth.
These are different goals, and they may sometimes conflict. A short answer can satisfy a user's immediate need while reducing the likelihood of a click.
Recognizing that trade-off is an important part of designing responsible information systems.
8. What Can Bloggers and Developers Do About It?
The shift toward AI-generated answers creates challenges, but it also gives creators a reason to rethink how they publish information.
If your work simply repeats information that already appears on hundreds of websites, it may be difficult to distinguish your article from a generated summary.
Originality becomes more important.
For bloggers and developers, a few strategies are worth considering.
Publish firsthand experience. Share what you built, what failed, what you measured, and what you learned. Real experiments provide context that generic explanations often lack.
Show your evidence. Include source links, code, screenshots, test results, and relevant examples. Make it possible for readers to verify your claims.
Explain the reasoning. Do not just provide the final solution. Explain why it works, when it might fail, and what alternatives you considered.
Build direct relationships with readers. Newsletters, communities, open-source projects, and professional networks can help people discover your work without depending entirely on search traffic.
Make content genuinely useful. Clear documentation, practical tutorials, original research, and detailed troubleshooting guides serve readers who need more than a quick answer.
None of these strategies guarantees traffic. Search engines, recommendation systems, and audience preferences will continue to change. But publishing work with distinctive value gives readers a reason to seek it out.
Conclusion: Answers Are Useful, but Discovery Still Matters
The internet is moving toward a model in which people can ask questions and receive immediate answers without navigating multiple websites.
That change offers genuine benefits. It can reduce friction, make complicated subjects easier to approach, and help people find information more efficiently.
But the internet is not simply a database of facts waiting to be summarized.
It is also a collection of people, communities, experiments, arguments, documentation, and original discoveries. Websites are not just containers for answers. They are places where knowledge is created, tested, challenged, and improved.
If AI systems make information easier to consume while making original sources harder to sustain, we could end up with a more convenient internet that has fewer incentives to produce the detailed work on which its answers depend.
That outcome is not inevitable. Search systems can link to sources, publishers can develop more distinctive resources, and readers can continue exploring beyond the first answer.
As developers, we should care about this because the open web has always been one of our most valuable learning resources.
The next time an AI system answers a technical question, consider opening the documentation or article behind it. Read the original explanation. Explore the alternatives. Support the people who took the time to investigate the problem.
The future of the internet should not be measured only by how quickly we receive an answer, but also by whether the people creating useful knowledge can continue doing so.
Because when we lose the habit of discovering sources, we risk losing more than clicks. We risk weakening the ecosystem that makes reliable answers possible in the first place.
Top comments (0)