We are halfway through 2026 and I need a minute to just take stock. Artificial intelligence has been one of the central topics of this year so far, and nothing makes that clearer than what happened just days ago. Anthropic, an AI startup that is only five years old, raised $65 billion in a single funding round, pushing its valuation to $965 billion. That put it ahead of OpenAI, previously the most valuable AI company in the world at $852 billion, and above Walmart, one of the largest and most established retailers on the planet, which sits at a market cap of $948 billion.
A company with no stores, no supply chain, and a fraction of the workforce of a traditional giant is now being valued higher than all of them. That is the world we are stepping into in the second half of 2026.
The First Half of 2026 Was Unlike Anything Before It
If you felt like AI news was coming at you from every direction this year, you were not imagining it. The pace of development in the first six months of 2026 has been genuinely without precedent.
Model releases came faster than most people could keep up with. Every few weeks, a new version, a new capability, a new benchmark broken. OpenAI, Google, Anthropic, Meta and others were all shipping updates at a speed that made it hard to even remember what the previous version could and could not do.
AI also started connecting to the tools and apps people already use, turning a chat window into something that could actually take action in the real world. That shift, quiet as it was, is one of the most significant things that happened this year.
And yet despite all of this, most people are still experiencing AI the same way they did two years ago. Opening a tab, typing a question, getting an answer, closing the tab. The tools have moved. The habits largely have not. That gap is what I find most interesting about this moment.
What AI Can Actually Do in 2026
The real question at this point is probably not what AI can do, but what it cannot. The capabilities keep growing. It is the human adoption of those capabilities that is lagging behind.
It can see, hear and read almost anything. You can take a photo of a confusing bill, a rash on your skin, a contract you do not understand, or a meal you want to recreate, and AI will engage with it meaningfully. Upload a PDF, paste a YouTube link, record a voice note. Tools like ChatGPT, Claude and Gemini can now process all of these in a single conversation without switching apps.
It can take actions on your behalf. AI is no longer limited to generating text and waiting for you to do something with it. It can now browse the web, fill out forms, summarise emails, manage files and complete multi-step tasks with minimal input from you. This is what people mean when they talk about AI agents. You give it a goal and it works toward it. You can access early versions of this today through ChatGPT, Claude and Gemini.
It can help you think through health questions more seriously. This is not about replacing your doctor. But AI has become a genuinely useful thinking partner for understanding a diagnosis, preparing questions before an appointment, or making sense of medical language that would otherwise send you down a confusing Google spiral. People are walking into consultations better informed than ever because they spent twenty minutes with an AI tool the night before.
You can have your own personal AI assistant. Beyond the general chat tools, people are now setting up dedicated AI assistants with their own names, personalities and areas of focus, accessible via Telegram, WhatsApp or a browser. Tools like Hermes and OpenClaw are examples of what that looks like in practice. It sounds technical but the barrier to getting one set up is lower than most people think.
It can generate video from a text description. Describe a scene in plain language and AI will produce a video clip from it. No camera, no editing software, no production budget. This capability has matured significantly in 2026 and is now accessible through several consumer facing tools without any technical background required.
It can build things for you, no coding required. Ordinary people are now describing what they want in plain language and watching AI produce working apps, simple websites, spreadsheet tools and automations. It is not perfect and it takes some patience, but the barrier to building something that actually works has never been lower. Claude and ChatGPT both support this directly in the browser.
This is by no means a complete list. AI capabilities are growing at a pace that is genuinely hard to keep up with, and that is actually one of the reasons I started this blog. To document it, share it, and make sense of it alongside you. Even then, some weeks it still feels like drinking from a fire hose.
What I Am Building Toward
I have been thinking a lot about this gap between what AI can do and how most people are actually experiencing it, and it is shaping everything I want to do in the second half of this year.
Since I started this blog I’ve been thinking of putting together a fun interactive workshop for people with little or no technical background who want to get hands-on with AI. Not theory. Actual practice, things like setting up your own AI assistant, learning to vibe code your first simple websote, and understanding what is worth paying attention to in a space that moves this fast. That is still in the planning phase but should be coming up with a date soon. Subscribe to the August Wheel newsletter to stay updated.
I will also be sharing interesting events and learning opportunities as I come across them, because part of staying ahead in this space is knowing where the conversations are happening.
If that sounds like something you want to be part of, the best place to stay in the loop is the newsletter. I keep it practical, honest, and free of the hype that tends to cloud this space.
FAQ
How can I use AI for more than just answering questions?
Start by giving AI context about what you are working on rather than just asking isolated questions. Use it to think through decisions, generate first drafts, or organise your thoughts. The more context you provide, the more useful the output becomes. Tools like Claude and ChatGPT are capable of sustained, multi-turn conversations that build on each other, which is very different from a one-question search.
What is the difference between using AI as a chatbot versus using it as an assistant?
A chatbot interaction is typically transactional: you ask, it answers, you leave. Using AI as an assistant means treating it as an ongoing collaborator. You give it background, you iterate, you ask follow-up questions, and you build toward something. The technology is the same. The approach is what changes the outcome.
Why is AI adoption still relatively low even though AI tools are everywhere?
Awareness and access are not the same as adoption. Most people have heard of ChatGPT and many have tried it once or twice. But building a regular habit around AI requires understanding what it is actually good for, and that understanding is still not widespread. The tools have outpaced the education, which is part of why workshops and practical resources matter right now.
The Second Half Starts Now
Six months ago, AI was already moving fast. Today it is moving faster, and the companies building it are being valued accordingly. But the most interesting question is not about valuations or model benchmarks. It is about what ordinary people do with these tools.
The future is here. It is just not evenly distributed yet. And that is exactly what the next six months are about.
Leave a Reply