Most people have asked ChatGPT a health question. Most of them had no idea that according to a National Institute of Health study, general AI models including ChatGPT, Gemini, and Grok have anywhere between a 50 to 80 percent hallucination rate on medical questions. That is not a small margin of error when your health is on the line.

In this episode of Real World AI Talks by Just AI News, host and editor-in-chief Paulo Iñaki Palma sits down with Benjamin Stover, Chief Commercial Officer of AI Med Consult, a specialist medical AI platform built specifically for cosmetic and plastic surgery clinics. The conversation goes deep into one of the most important and least discussed distinctions in AI right now: the difference between generalist AI and specialist medical AI, and why that difference matters more than most people realize.

In this episode we cover:

• Why ChatGPT and other general AI models have a 50 to 80 percent hallucination rate on medical questions according to a NIH study
• Why you should never rely on a general chatbot for medical advice and what to use instead
• What a medically trained AI actually means in plain English
• How AI Med Consult was built on medical textbooks and refined by real practicing surgeons over two and a half years
• Why the AI does not crawl the internet like ChatGPT and why that matters for accuracy
• How the platform gets fully customized for each individual clinic in about a week
• Whether OpenAI or Anthropic could ever compete in specialist medical AI
• How AI is being used to engage patients at two in the morning when nobody at the clinic is available
• The difference between AI that gives medical advice and AI that qualifies and converts patient leads
• Why businesses that start using AI now will have a competitive advantage that late adopters will struggle to overcome
• Whether the future belongs to generalist AI or specialist AI and why the answer is both
• What will define a trustworthy medical AI company over the next five years
• Where AI Med Consult is expanding next

AI is not replacing doctors. But it is changing how clinics find, engage, and convert patients. This episode explains what that actually looks like in practice, built on real clinical knowledge, tested by real surgeons, and designed to handle the conversations that happen when nobody at the clinic is picking up the phone.

Real World AI Talks is hosted by Paulo Iñaki Palma, editor-in-chief of Just AI News, and is available on Spotify, Apple Podcasts, and YouTube. New episodes every week.

What is agentic commerce? Most people have never heard the term. But agentic commerce has arrived, and it is already changing how people buy things, run businesses, and make payments online. In this episode of Real World AI Talks by Just AI News, host and editor-in-chief Paulo Iñaki Palma sits down with Benjamin de Waal, CTO, and David Parkinson, CEO of Musqet, to break down the future of agentic payments and what it means for merchants, businesses, and everyday consumers right now.

Agentic commerce means your AI agent handles the entire purchase for you. It finds the product, reads the store data, adds items to the cart, and completes the checkout. Want to know how to buy things with your Bitcoin without touching a website? This episode explains exactly how that already works using Bitcoin Lightning payments and the UCP protocol. Ben de Waal walks through how to build agentic commerce right now, what makes it work reliably, and why letting your AI agent pay with Bitcoin is safer than most people assume.

The conversation also goes somewhere most discussions avoid. AI agents are paying each other now. Not in theory. In practice. David Parkinson says we are within twelve months of fully AI led, AI run, AI financed companies that still employ humans for physical tasks. Ben de Waal adds that the question of which decisions AI should never make has a much less reassuring answer than most people expect.

In this episode we cover:

• What is agentic commerce and how it works in practice today
• How to buy things with your Bitcoin using an AI agent
• Why agentic commerce has arrived and what that means for merchants
• The future of agentic payments and where it is heading
• How to build agentic commerce into your own business or platform
• Why AI agents are paying each other now and what that means
• How Musqet built a full agentic checkout with Bitcoin Lightning payments
• Why the UCP protocol matters for agentic commerce
• How much of a business AI can run without human input
• What fully AI led, AI run, AI financed companies will look like
• Why being a small company is an advantage when AI is moving this fast
• How AI avatars with real celebrity personalities are changing customer interaction
• Why Stripe is struggling with agentic commerce and what smaller companies are doing differently

This is one of the most practical and honest conversations about agentic commerce, Bitcoin payments, and AI business operations we have published. If you are building in this space, working in payments, or simply want to understand where commerce is heading, this episode gives you real answers from two founders doing it right now.

Real World AI Talks by Just AI News is hosted by Paulo Iñaki Palma, and is available on Spotify, Apple Podcasts, and YouTube. New episodes every week.

Big data and AI are on a collision course. And most companies are not ready for what comes next.

In this episode of Real World AI Talks by Just AI News, host and editor-in-chief Paulo Iñaki Palma sits down with Tom Howe, Director of Insights Engineering at Hydrolix, a real time data analytics platform trusted by companies from growing e-commerce businesses all the way up to Fox, whose Super Bowl stream Hydrolix helped power in real time. The conversation covers one of the most underrated problems behind the AI boom: how companies are collecting petabytes of data, spending millions on AI tools, and have no clear strategy for what to do with any of it.

Most companies think more data means better AI. Tom Howe says that is one of the most expensive misconceptions in the industry right now. Without a clear AI data strategy, companies end up in a feedback loop where AI is analyzing AI, producing confident answers based on wrong or incomplete data, and nobody is catching it. Tom has seen entire data clusters crash because a poorly configured LLM wrote queries that took down the hardware entirely.

In this episode we cover:

• Why companies are collecting more data than they can actually use
• What happens when an LLM works with poorly structured or incomplete data
• Why AI giving you a confident answer is not the same as a correct one
• How to decide what data is useful and what is just expensive noise
• What AI bots are doing to web traffic and why blocking them all is the wrong move
• How generative engine optimization is replacing traditional SEO
• Why the dependency on advertising revenue is about to become a serious business problem
• What major changes to internet trust and security could look like in the next five years
• Why open source AI models are changing the data infrastructure conversation
• What real time data analytics actually looks like on Super Bowl day

AI is not just a productivity tool. It is generating more data than most companies know what to do with. Tom Howe has spent his career solving exactly that problem, and this is one of the most practical and honest conversations about big data and AI infrastructure we have published.

Real World AI Talks is hosted by Paulo Iñaki Palma, editor-in-chief of Just AI News, and is available on Spotify, Apple Podcasts, and YouTube. New episodes every week.