AI with Kyle Daily Update 042

Today in AI: Anthropic larger than Disney + Fantasy authors in trouble?

Final Live Webinar: Learn How to Make $1,000-$4,000 an Hour Teaching Businesses Around the World How to Use AI.

📅 Thursday  Check link for timezone 🎟️ Free

The skinny on what's happening in AI - straight from the previous live session:

Highlights

🚀 Anthropic Raises £13B, Now Worth More Than Nike

Anthropic just closed a Series F funding round worth $13 billion, bringing their post-money valuation to $183 billion! To put that in perspective: they're now worth more than Boeing ($120B), Nike ($160B), Ford ($40B), GM ($50B), and even Disney ($170B). Not bad for a company that's barely three years old.

Kyle's take: The money being thrown at AI companies is mindboggling, which is why it’s helpful to compare to “traditional” companies. We're seeing valuations that dwarf century-old industrial giants, all based on future earnings that don't really exist yet…

That's potential bubble territory, and we saw a taste of that fear last week when markets wobbled over Palantir's inability to justify their valuation with actual revenue. That said, I'm delighted Anthropic got this funding - Claude is brilliant, and they're one of the few companies building AI responsibly.

📊 That Viral "Where AI Gets Its Facts" Chart Is Misleading Nonsense

A chart claiming "40% of AI facts come from Reddit" has been doing the rounds, but it's rubbish. The chart actually shows where LLMs go when they do web searches for supplemental information - not where they get their core knowledge from. That comes from pre-training data. The difference is massive: one is real-time search results, the other is the fundamental knowledge baked into the model.

Kyle's take: This is exactly the sort of chart that looks impressive but falls apart the moment you scratch below the surface. People are already using it to justify focusing on Reddit for business visibility, which is incorrect.

The real worry isn't just the misinformation - it's that we're losing our critical thinking skills. That's how we end up believing charts like this without questioning them. And then spreading them willy-nilly online without thought.

✍️ AI Writing Beats Fantasy Authors in Blind Test

Author Mark Lawrence ran a fascinating experiment with his readers. He mixed eight 350-word fantasy stories - some written by established authors (including Robin Hobb), others by GPT-5. Out of 964 voters, people only correctly identified three stories' origins, got three completely wrong, and couldn't decide on two others. Some readers thought Robin Hobb's actual writing was AI-generated, while AI stories were deemed human-written.

Kyle's take: This is unsettling if you're a writer. We're not talking about amateur authors here - Robin Hobb is a massive name in fantasy. The fact that 50% of people can't tell the difference between her work and GPT-5 suggests we're edging past the point where AI writing is indistinguishable from human creativity.

Combined with that Wired story from last week where they accidentally published an AI-written article, it's clear we need to have a serious conversation about authorship and authenticity. Because if we can't tell the difference, do we actually care who wrote it?

Member Question from Teva: "Have you noticed law firms now advising companies to update their TOS regarding company's use of AI?"

Kyle's response: Not specifically, but that makes complete sense. Legals always lag behind technology, but they're starting to catch up with AI implications. Companies definitely need to update their terms of service for the AI age - things like data scraping, agent access, user-generated AI content.

There's actually a brilliant business opportunity here: build a tool that scans existing ToS, flags AI-related gaps, then either generates updated versions or refers to trusted legal partners. The scanner could be free (great lead magnet!), paid updates could be £50-200, and human legal review could be thousands.

Go build it!!

Discussed at 26:50

Want the full unfiltered discussion? Join me tomorrow for the daily AI news live stream where we dig into the stories and you can ask questions directly.

Streaming on YouTube (with screen share) and TikTok (follow and turn on notifications for Live Notification).

Audio Podcast on iTunes and Spotify.