Skip to main content
Accidental Tech Podcast

709: Steel-Cable Insurance

127 min episode · 3 min read

Episode

127 min

Read time

3 min

Topics

Fundraising & VC, Leadership, Design & UX

AI-Generated Summary

Key Takeaways

  • A20 Pro single-core performance: The A20 Pro chip in the iPhone 18 Pro scores 24% better in single-core CPU performance than the A19 Pro and 11% better than the M5 iPad Pro, previously the single-core leader. GPU metal benchmark scores came in at 38% above the A19 Pro, aligning precisely with Apple's stated 40% gaming performance claim. This makes the A20 Pro the strongest single-core performer across Apple's entire current lineup, including desktop chips.
  • M6 chip generational gains: Preliminary Geekbench data shows the M6 delivers 10% better single-core and 35% better multicore CPU performance over the M5, with GPU scores 25% higher. These are unusually large year-over-year gains for a non-suffix chip. When comparing the M5 Ultra to the M5 Max, the Ultra shows roughly 60% better multicore performance, though benchmark parallelism limitations mean real-world rendering workloads may tell a different story than synthetic scores.
  • Variable aperture equivalents on iPhone 18 Pro: The iPhone 18 Pro's widest aperture of f/1.48 translates to approximately f/5.2 on a full-frame sensor, while the narrowest f/4 setting equals roughly f/14 full-frame equivalent. Apple's camera software defaults to f/1.8 for most shots, reserving the wider f/1.48 for low-light conditions. Third-party apps can access the full continuous aperture range beyond the four presets Apple exposes natively, including smooth aperture transitions during video recording.
  • Straight aperture blades and starburst rendering: The iPhone 18 Pro uses straight-edged aperture blades rather than rounded ones, a deliberate optical choice. Straight blades produce sharper, more defined starburst effects on point light sources, while rounded blades produce softer, less defined starbursts. The number of starburst rays equals twice the number of aperture blades. Stopping down to a higher aperture value in mixed artificial lighting also reduces sensor banding caused by LED flicker during video capture.
  • iPhone Duo developer tooling gap: As of recording, Xcode 26.1 beta — which includes the iPhone Duo simulator — has not yet shipped, despite the device launching imminently. Xcode 26.2 beta released instead, explicitly noting it does not include Duo development support. This leaves developers unable to test or adapt apps for the new form factor before customers receive devices. A parallel is drawn to Apple's past policy change blocking App Store reviews from beta iOS builds, suggesting a similar review moratorium for Duo apps would be fair.

What It Covers

ATP episode 709 covers the iPhone 18 Pro and iPhone Duo hardware analysis, including A20 Pro and M6 chip benchmark comparisons, variable aperture camera mechanics, foldable device usability tradeoffs, Apple Pencil compatibility gaps on the Duo, iOS 27 feature omissions, John Turnus's leadership style as new Apple CEO, and a St. Jude Children's Research Hospital fundraising campaign for September.

Key Questions Answered

  • A20 Pro single-core performance: The A20 Pro chip in the iPhone 18 Pro scores 24% better in single-core CPU performance than the A19 Pro and 11% better than the M5 iPad Pro, previously the single-core leader. GPU metal benchmark scores came in at 38% above the A19 Pro, aligning precisely with Apple's stated 40% gaming performance claim. This makes the A20 Pro the strongest single-core performer across Apple's entire current lineup, including desktop chips.
  • M6 chip generational gains: Preliminary Geekbench data shows the M6 delivers 10% better single-core and 35% better multicore CPU performance over the M5, with GPU scores 25% higher. These are unusually large year-over-year gains for a non-suffix chip. When comparing the M5 Ultra to the M5 Max, the Ultra shows roughly 60% better multicore performance, though benchmark parallelism limitations mean real-world rendering workloads may tell a different story than synthetic scores.
  • Variable aperture equivalents on iPhone 18 Pro: The iPhone 18 Pro's widest aperture of f/1.48 translates to approximately f/5.2 on a full-frame sensor, while the narrowest f/4 setting equals roughly f/14 full-frame equivalent. Apple's camera software defaults to f/1.8 for most shots, reserving the wider f/1.48 for low-light conditions. Third-party apps can access the full continuous aperture range beyond the four presets Apple exposes natively, including smooth aperture transitions during video recording.
  • Straight aperture blades and starburst rendering: The iPhone 18 Pro uses straight-edged aperture blades rather than rounded ones, a deliberate optical choice. Straight blades produce sharper, more defined starburst effects on point light sources, while rounded blades produce softer, less defined starbursts. The number of starburst rays equals twice the number of aperture blades. Stopping down to a higher aperture value in mixed artificial lighting also reduces sensor banding caused by LED flicker during video capture.
  • iPhone Duo developer tooling gap: As of recording, Xcode 26.1 beta — which includes the iPhone Duo simulator — has not yet shipped, despite the device launching imminently. Xcode 26.2 beta released instead, explicitly noting it does not include Duo development support. This leaves developers unable to test or adapt apps for the new form factor before customers receive devices. A parallel is drawn to Apple's past policy change blocking App Store reviews from beta iOS builds, suggesting a similar review moratorium for Duo apps would be fair.
  • iPhone Duo MagSafe and accessory compatibility: The Duo's short, wide aspect ratio means many existing MagSafe accessories — including wallets and battery packs — will either extend beyond the device's edges or sit atop the camera plateau, preventing flat contact. The magnetic alignment circle may fit, but the physical footprint of standard MagSafe accessories was designed around taller iPhone proportions. Buyers should verify Duo-specific compatibility before purchasing accessories rather than assuming standard MagSafe labeling guarantees a usable fit.
  • Apple Pencil support on iPhone Duo is limited: The iPhone Duo supports only the USB-C Apple Pencil, which lacks pressure sensitivity, not the Apple Pencil Pro. The Pencil Pro requires magnetic attachment for both pairing and charging, and the Duo provides neither. A rumored bespoke pressure-sensitive pencil designed for the Duo was reportedly canceled during development, with the USB-C model serving as a fallback. Third-party stylus makers are positioned to fill this gap with Duo-optimized pressure-sensitive options.

Notable Moment

A CarMax listing revealed that a 2018 Volkswagen Golf R purchased new for roughly $37,000 was being resold in 2026 for nearly $39,000 — more than the original purchase price eight years later. The seller had already accepted $26,000 from CarMax, illustrating how rare-trim vehicles can appreciate in ways that make conventional depreciation assumptions unreliable for niche enthusiast models.

Know someone who'd find this useful?

Episode Transcript

I had been looking to see whether or not my VIN number for my Volkswagen had been listed online for sale, and it hadn't been. I've been checking every other day or so because I'm me, and I like to know these things. Why? Are you stalking your ex? Because I'm curious. There is no good no good can come from this. It's just find it interesting, and I was curious what they were going to list it for. And so But but how okay. There's only one way this can go. Like, they're not gonna list it for below what they paid you. Oh, of course. Course. So they're gonna list it for some amount of money that is above what they paid you. Mhmm. And the only thing that can possibly do is make you feel bad that they didn't pay you enough. Well, so on paper, especially when we're talking about me, yes. However, however, first of all, I'm needy and I just want to know, and second of all, I I feel like I had done the best I could do short of a private sale. Right? I had surveyed a couple of different options, I kinda sorta and it was my own choice, but I kinda sorta wanted to do it with a quickness, and I had and I had done what I thought was best for me at the time, and I stand by that. Now if I had sold it privately, I'm sure I could have gotten quite a bit more money, but I didn't want to mess with any of that. I have no patience for that. I absolutely did not want to, have somebody test drive the car whether or not I'm in it, but watching them like melt my clutch, no thank you. Uh-uh. So I did what I think was right for me. Yeah. The whole reason to sell to a like a company like that is so that it's out of your hands. It's quick, it's done, like you're not getting the best possible price, but you're at least like not having to deal with any of that. Exactly. The last thing you should possibly then know is, like, so how much profit are they gonna make off of me? I feel like there's just there's no good that comes from that. Well, I again, I think it's fascinating just because I was curious what they would do, and I when I sold the BMW to a local, like, CarMax equivalent, it was one of those, we'll beat CarMax prices, but they beat it by literally a $150, I believe. Anyways, when I sold it to them, I forget what they listed it for, but it was, I don't know, a couple grand more than they paid me, which is what I expected. Right? I mean, was a relatively unremarkable three series, and it was relatively high miles, and they sold it presumably for a little bit more …

Get the full transcript (27,420 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all Accidental Tech Podcast transcripts →

You just read a 3-minute summary of a 124-minute episode.

Get Accidental Tech Podcast summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

Books, tools, and gear mentioned in this episode

SignalCast may earn commission on purchases via these links. As an Amazon Associate, SignalCast earns from qualifying purchases.

Tools

  • by Anthropic

    Sponsors include Claude at https://claude.ai/atp
  • Sponsors include Granola at https://granola.ai/atp
  • Preliminary Geekbench data shows the M6 delivers 10% better single-core and 35% better multicore CPU performance over the M5, with GPU scores 25% higher.
  • by Apple

    As of recording, Xcode 26.1 beta — which includes the iPhone Duo simulator — has not yet shipped, despite the device launching imminently. Xcode 26.2 beta released instead, explicitly noting it does not include Duo development support.
  • Sponsors include ZocDoc at https://zocdoc.com/atp

Gear

  • by Apple

    ATP episode 709 covers the iPhone 18 Pro and iPhone Duo hardware analysis, including A20 Pro and M6 chip benchmark comparisons, variable aperture camera mechanics, foldable device usability tradeoffs
  • by Apple

    The iPhone Duo supports only the USB-C Apple Pencil, which lacks pressure sensitivity, not the Apple Pencil Pro.
  • by Apple

    ATP episode 709 covers the iPhone 18 Pro and iPhone Duo hardware analysis, including A20 Pro and M6 chip benchmark comparisons, variable aperture camera mechanics
  • by Apple

    The iPhone Duo supports only the USB-C Apple Pencil, which lacks pressure sensitivity, not the Apple Pencil Pro. The Pencil Pro requires magnetic attachment for both pairing and charging

More from Accidental Tech Podcast

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best Tech Podcasts (2026) — ranked and reviewed with AI summaries.

You're clearly into Accidental Tech Podcast.

Every Monday, we deliver AI summaries of the latest episodes from Accidental Tech Podcast and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime