Skip to content
IOS 27 Beta Time episode artwork

Jul 19, 2026 51 min

IOS 27 Beta Time

from The Insanely Great Podcast Hosted by Christopher John

This week Christopher and DJ trade first impressions of the iOS 27 public beta, then wander (as usual) into everything from Apple's courtroom drama with OpenAI to the right way to wear sunblock in 115-degree heat. In this episode: First impressions of iOS 27 public beta 3 — snappier performance, and where advanced dictation settings are (or aren't) hiding A weird predictive-text bug that shows up right before pasting text Whether Apple is intentionally gating advanced AI features to newer hardware — and why Apple's Gemini partnership might not be permanent LLMs vs. "world models" — and why current models can't really form opinions The Apple/OpenAI lawsuit: a former Apple hardware engineer, a laptop full of trade secrets, and why Apple named OpenAI but not Jony Ive or Love From Why Anthropic and Apple's rumored Perplexity acquisition talks reportedly fell apart Siri's new "Ask Siri or Ask ChatGPT" toggle, and Apple's Optimized Search & Siri indexing Apple's odd content filtering in Photos search (yes, really — "bikini" vs. "swimsuit") Third-party dictation tools compared: Grammarly Pro, Whisperflow, MacWhisper, and Voice Notes How the show's own transcription and AI-assisted show notes workflow works Chris's annual "golf trip to hell" in Mesquite, Nevada — and a genuinely deep dive into chemical vs. oxide sunblock What's still missing from the betas (RCS encryption) and why Samsung gave up on its own messaging app New AI photo editing tools in iOS 27 — object removal, perspective shifts, and "Nano Banana" Kimi K3, the largest open frontier model yet, and how it stacks up against Grok, Claude/Fable, and Gemini Anthropic's Fable model helping Microsoft triage a record-breaking Patch Tuesday Using Fable/Claude to redesign a website and get real performance/security recommendations Why Anthropic's models hallucinate less but feel more "conservative" — and why that's driving tools like Claude Code Gemini 3.5 Pro's delayed coding rollout, and Kimi K3's surprisingly high price tag Chapter Markers (00:00) - – Intro (00:33) - – iOS 27 beta performance impressions (01:40) - – Predictive text bug when pasting (02:03) - – Hunting for the advanced dictation toggle (02:54) - – Hardware gating: which iPhones get advanced dictation (03:00) - – Is Apple intentionally holding back AI features? (06:31) - – LLMs vs. "world models" and self-learning AI (07:01) - – Can you give an AI an opinion/personality? (07:40) - – Apple's rumored strategy: buying AI companies vs. poaching talent (08:38) - – Why the Apple–Perplexity acquisition talks fell apart (09:03) - – Is Siri slower on the betas? (09:20) - – New "Ask Siri or Ask ChatGPT" toggle (09:45) - – Apple vs. OpenAI lawsuit: the Tan/Jony Ive/Love From backstory (14:00) - – Anthropic vs. OpenAI's government stance & ARR growth (14:42) - – Optimized Search & Siri indexing, battery life gains (16:00) - – Photos indexing differences across devices (19:11) - – Apple's odd search-term filtering ("bikini" vs. "swimsuit") (20:32) - – Proofreading shortcuts and vulgarity filters (21:04) - – Third-party dictation apps: Grammarly Pro, Whisperflow (22:23) - – MacWhisper and Voice Notes compared (23:20) - – How the show's transcription & AI show-notes workflow works (25:15) - – The annual golf trip to Mesquite, Nevada (27:40) - – Sunblock deep dive: chemical vs. oxide, sunblock drinks (31:21) - – What's expected in beta 4 (32:19) - – Features still missing from the betas (RCS encryption) (34:09) - – Why Samsung gave up on its own messaging app (35:44) - – New AI photo editing tools in Photos (object removal, perspective shift, "Nano Banana") (37:30) - – Siri's improved sports scores & phone lookups (38:51) - – Simplified beta channels — no dev account needed (39:00) - – Apple Watch's Siri app and its phone dependency (41:11) - – Photo search improvements (42:06) - – Kimi K3: the largest open frontier model (42:59) - – Grok's parameter race and photo-editing use (44:07) - – Microsoft's record Patch Tuesday & Fable's role in triage (44:59) - – Fable (Claude) export-control situation resolved (45:06) - – Using Fable to redesign a website (46:35) - – Why Anthropic's models hallucinate less but feel "conservative" (47:54) - – Claude Code and WordPress integration (49:16) - – Gemini 3.5 Pro's delayed coding rollout (49:53) - – Kimi K3 pricing vs. DeepSeek (50:54) - – Wrap-up, next episode timing, listener growth

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

IOS 27 Beta Time

0:00 51:53

Chapters

  1. 0:00 – Intro
  2. 0:33 – iOS 27 beta performance impressions
  3. 1:40 – Predictive text bug when pasting
  4. 2:03 – Hunting for the advanced dictation toggle
  5. 2:54 – Hardware gating: which iPhones get advanced dictation
  6. 3:00 – Is Apple intentionally holding back AI features?
  7. 6:31 – LLMs vs. "world models" and self-learning AI
  8. 7:01 – Can you give an AI an opinion/personality?
  9. 7:40 – Apple's rumored strategy: buying AI companies vs. poaching talent
  10. 8:38 – Why the Apple–Perplexity acquisition talks fell apart
  11. 9:03 – Is Siri slower on the betas?
  12. 9:20 – New "Ask Siri or Ask ChatGPT" toggle
  13. 9:45 – Apple vs. OpenAI lawsuit: the Tan/Jony Ive/Love From backstory
  14. 14:00 – Anthropic vs. OpenAI's government stance & ARR growth
  15. 14:42 – Optimized Search & Siri indexing, battery life gains
  16. 16:00 – Photos indexing differences across devices
  17. 19:11 – Apple's odd search-term filtering ("bikini" vs. "swimsuit")
  18. 20:32 – Proofreading shortcuts and vulgarity filters
  19. 21:04 – Third-party dictation apps: Grammarly Pro, Whisperflow
  20. 22:23 – MacWhisper and Voice Notes compared
  21. 23:20 – How the show's transcription & AI show-notes workflow works
  22. 25:15 – The annual golf trip to Mesquite, Nevada
  23. 27:40 – Sunblock deep dive: chemical vs. oxide, sunblock drinks
  24. 31:21 – What's expected in beta 4
  25. 32:19 – Features still missing from the betas (RCS encryption)
  26. 34:09 – Why Samsung gave up on its own messaging app
  27. 35:44 – New AI photo editing tools in Photos (object removal, perspective shift, "Nano Banana")
  28. 37:30 – Siri's improved sports scores & phone lookups
  29. 38:51 – Simplified beta channels — no dev account needed
  30. 39:00 – Apple Watch's Siri app and its phone dependency
  31. 41:11 – Photo search improvements
  32. 42:06 – Kimi K3: the largest open frontier model
  33. 42:59 – Grok's parameter race and photo-editing use
  34. 44:07 – Microsoft's record Patch Tuesday & Fable's role in triage
  35. 44:59 – Fable (Claude) export-control situation resolved
  36. 45:06 – Using Fable to redesign a website
  37. 46:35 – Why Anthropic's models hallucinate less but feel "conservative"
  38. 47:54 – Claude Code and WordPress integration
  39. 49:16 – Gemini 3.5 Pro's delayed coding rollout
  40. 49:53 – Kimi K3 pricing vs. DeepSeek
  41. 50:54 – Wrap-up, next episode timing, listener growth

No transcript yet.

Requests help us decide what to transcribe next.

Frequently asked questions

How long is this episode of The Insanely Great Podcast?

This episode is 51 minutes long.

When was this The Insanely Great Podcast episode published?

This episode was published on July 19, 2026.

Can I download this The Insanely Great Podcast episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.