Podcast Summary: "Do We Need AI Browsers? What Are Claude Skills?"

This Day in AI Podcast (EP99.22) — October 24, 2025
Hosts: Michael Sharkey & Chris Sharkey

Episode Overview

Michael and Chris Sharkey dive into the week's new crop of AI browser announcements, specifically ChatGPT Atlas and the Microsoft Copilot AI browser. They debate the necessity (or lack thereof) of AI-integrated browsers, critique their real-world usefulness, and discuss broader trends in AI’s relentless march into every digital corner. The episode’s latter half explores a feature Anthropic launched—Claude Skills—and why it might actually represent meaningful progress compared to the cosmetic "innovation" of browser sidebars.

The vibe throughout is humorous, candidly exasperated, and a little irreverent: “We were promised AGI and we got a web browser.”

Key Discussion Points & Insights

1. The "AI Browser" Trend: ChatGPT Atlas and Microsoft Copilot

• AI Browsers as Repurposed Features
  Chris compares the never-ending tweaks in AI software to Marge’s endlessly-altered dress in The Simpsons—different superficial covers, fundamentally the same underneath.
  • "We've got sort of the ChatGPT, the chat paradigm, if you will. And then it's getting repackaged, repurposed up." (Chris, 00:07)
• ChatGPT Atlas:
  • A Chrome-based browser, Mac-only at launch (“makes sense when Copilot announced their own version of this exact same browser only a day later.”)
  • Main feature: an integrated chat sidebar that acts in the context of the current page.
  • OpenAI’s examples are underwhelming: shopping lists, booking travel, beach day reminders.
• Lack of Practical Utility
  Chris tried using Atlas for an actual task:
  • “I think it spun for like nine or ten minutes trying to click around and operate like one control on the website...I just didn’t find it very useful.” (Chris, 02:40)
  • He predicts people will treat these browsers as tech demos, not tools for work.
• Browser Should Be a Neutral Shell
  Chris and Alex (guest) discuss the fundamental expectation that browsers shouldn't be all-absorbing platforms:
  • “A browser should be a neutral shell, like a window to the Internet and the Internet is the applications…Do you really always want them watching and storing memories?” (Chris, 03:32)
  • Concerns about privacy, prompt injection, safety, and unnecessary AI onboarding.

2. Surveillance & Censorship Concerns

• AI Deciding What You Can See
  Chris recounts an example where searching for Hitler videos returns a safety warning rather than actual results:
  • “You're now bringing a browser onto your computer that is, one, monitoring you and how you use the Internet…And then, two, now has the safety police of the model deciding what you can and cannot research or consume or do.” (Chris, 07:37)
• Not for Enterprise Use
  • “People in corporate environments…can never EVER use a browser like this…in terms of transmitting data…you’ve breached your rules.” (Alex, 08:05)
  • Security, compliance, and privacy are potentially compromised.

3. The Arms Race of Market Share

• Disguised Play for User Data
  • The hosts are skeptical about AI browsers as genuine innovations versus attempts to capture data and usage patterns.
  • “They literally just want to get access to everything you're looking at for whatever reason to like maybe add some value, maybe make it slightly better for you, but also get training data.” (Alex, 09:26)
• Distribution Over Usefulness
  • Chris: “They're basically bribing people that aren't paying for ChatGPT by giving them extra usage for making it the default browser at the system level.” (Chris, 15:37)
  • Equates it to “rich guy billionaire market share wars.”

4. Microsoft Copilot AI Browser & Humanist AI

• Nearly Identical to ChatGPT Atlas
  • Copilot’s Edge-based browser introduces similar chatbox-centered features after Atlas’s Mac launch.
  • “So wait, it looks exactly the same.” (Chris, 19:30)
• Marketing Gimmicks: “Humanist AI”
  • “All of today’s Copilot announcements boil down to one core idea: We’re betting on humanist AI. An AI that always puts humans first.” (Chris, 18:20)
  • The hosts ridicule the notion, seeing it as superficial.
  • “It’s just like marketing…We were trying to replace humans and now we’re not.” (Alex, 19:31)
• Frustration with Lack of Progress
  • Repeatedly, Chris and Alex express fatigue with “yet another chat box in a browser,” lamenting the lack of real newness.
  • “I just think a year later, like, come on, a year later and where are we? Like, I mean, not much has changed.” (Chris, 22:06)

5. Real AI Use Cases: Where the Value Actually Is

• Power of Contextual, Agentic Workflows
  • Chris outlines his real-world use: a custom MCP assistant in Simtheory that accesses support tickets, internal databases, Stripe, and documentation to automate responses.
  • “Every time I use it, I'm like this thing is magical...as it's responding to common things over time it's getting better.” (Chris, 25:30)
• Preference for Substance Over Skin-Deep Features
  • What’s truly exciting is exposing tools and data to the AI to revolutionize productivity—“absolutely incredible power that’s sitting there right now if a company…is able to expose it.”
  • Alex describes an example where an AI agent queries a database, interprets results, and produces solutions in minutes rather than days. (29:28)
• Desire for More OpenAI/Anthropic Tutorials
  • “A lot of people in the community have been asking for us to do some tutorials …and we plan to do it because we think it’s a different way of working and once you discover it it really is pretty game changing.” (Chris, 28:11)

6. Claude Skills: Actual Progress and Potential

• Transition Away from Browsers to Power User Paradigms
  • The episode’s final segment introduces “Claude Skills”: containers of instructions, scripts, and resources for task-specific enhancement.
  • Alex says the more he dives in, the more significant Claude Skills seem:
    • “It’s actually an incredibly powerful paradigm of working and it’s going to be a little bit tricky with the words like skills and tools and all these things to understand the difference. But there is a distinct difference and there are major direct advantages to… [TBC in next episode].” (Alex, 35:16)
• Optimism for Customizability and Productivity
  • Idea: Instead of manually guiding the AI through context building (the “foreplay” analogy), you can predefine that ramp-up in a reusable “skill.”
  • “How do you get the AI in that flow state straight away? So you sit down, you open your 16 chat windows or whatever and then you're just getting stuff done, launching tasks asynchronously, and being a power worker.” (Alex, 27:34)

Notable Quotes & Memorable Moments

• Simpsons Marge Dress Analogy
  “AI is being repurposed from that episode in the Simpsons where Marge has the dress…and then she keeps changing it up all the time, but it's really fundamentally the same dress. That's what it feels like is happening in AI right now.”
  Chris, 00:04

• On AI Overreach in Browsers
  “We were promised AGI and we got a web browser.”
  Chris, 10:43

• On Attempts to Win Market Share
  “It’s sort of like rich guy billionaire market share wars, where...they've got the risk board of the world and they're like, if we get this many billion people using this browser, then...we will control the world for the next little while.”
  Alex, 16:03

• Fatigue with Faux Product Announcements
  “If this was the forefront of AI, I'd quit this podcast immediately.”
  Alex, 22:19

• Pointlessness of Current Consumer AI Offerings “It feels like we're getting the sterilized, watered-down version because they think people are stupid...just dumb it down to just the most dumb possible, then maybe they'll use it.” Chris, 32:45

• On True AI Value
  “The second someone has that moment — they're like, I could use it for this, I could use it for that. This is going to change the thing. The AI just did work that would have taken me a week in 30 seconds.”
  Alex, 23:15

Timestamps of Key Segments

• 00:03 — Simpsons Dress Analogy; Overview of Atlas as “the same dressed-up chat paradigm”
• 02:00–04:50 — Atlas features, hands-on frustration, examples lack real utility
• 07:37 — Safety & censorship alarm bells: “I can’t browse or display videos of Adolf Hitler…”
• 09:26–10:42 — Market share grabs; skepticism of AI browser “innovation”
• 14:03–16:45 — Lowering the bar for mass adoption; “bribing users”
• 18:20–21:42 — Copilot’s “humanist AI;” poking fun at indistinguishable features
• 25:30–29:28 — Real-world AI agent workflows; SimTheory custom MCP example
• 33:26–34:13 — Frustration that vendors don’t showcase genuinely powerful workflows
• 35:16 — Claude Skills introduction and optimism

Conclusion

This episode is a candid, humor-laced critique of the “AI browser” movement, dismissing it as superficial and likely to fade as quickly as the plug-in frills of the early 2000s web. The Sharkey brothers are much more bullish on meaningful AI progress like customizable workflows and deep tool integration (see: Claude Skills), where the power-user can actually get magical results—and the average user just needs a better invitation, not a browser with more chatboxes.

For listeners wanting to understand why the browser wars are such a sideshow compared to what’s possible with today’s LLM agents, this episode is an excellent summary of both the frustration and optimism in the AI space.