Content About Stocktwits
No Image Available

Best Laid Plans: Apple’s Design, Europe’s AI, & Anthropic’s Blip 2.0. ARD #102

Author
Michael Parekh AI: Reset to Zero
21 hours ago
14 MIN READ

Today’s theme is “The Best Laid Plans of Mice and Men” — the immortal line from Robert Burns’ 1785 poem “To a Mouse,” reminding us that no matter how carefully one prepares, unforeseen events can almost always disrupt the initial goals. I want to dig into three events in the AI world that underline exactly that — carefully laid plans, by companies and by countries, going very differently than planned — for the AI Tech Wave, each with my Take first, then my Overall Take. Plus a Roomba Gadget AI and two questions that lean into the same theme.

(1) Apple’s Design Gauntlet Under New CEO John Ternus

MP TAKE: For over a decade, Apple has been laying out its design plans post-Jony Ive — its long-departed, iconic design head, who worked directly with Steve Jobs until Jobs passed away. The core thing to absorb: those long-term design plans now have to be redone — not undone, but redone — because of a host of unforeseen events. The biggest, of course, is AI.

Ive now works with OpenAI (the $6.5 billion acquisition of his AI-device startup), planning all manner of AI devices — potentially even an AI smartphone. Apple has taken other dents to its design bench, including a key designer hired away by Mark Zuckerberg’s Meta (Alan Dye). Design is the heart and soul of Apple — what we think of as MacBooks, iPhones and wearables — so the major to-do for incoming CEO John Ternus (who takes over this September as Tim Cook steps away after 15 years, having just warned of price increases from global supply constraints) is to reformulate these design plans and appoint leaders at the senior-most levels, presumably reporting to Ternus himself. Mark Gurman lays it out in detail at Bloomberg. I continue to think the Siri AI Apple demoed at WWDC could be the Apple device that becomes the AI device we’ve all been waiting for — but there’s a lot of software and hardware design work ahead, now under Ternus and Apple-silicon chief Johny Srouji, with a rich pipeline of foldables, camera-equipped AirPods and more. Apple thought it had this wired and figured out; now it has to evolve it — a telling case of best-laid plans meeting AI.

Sources, in narrative order: Bloomberg / Power On (Mark Gurman) — Apple’s New CEO Must Rebuild a Design Team That Lost Its Way. For longtime readers, in narrative order: new bounce for Apple design in the AI age in AI-RTZ #975; Meta amping up its AI device-design capabilities in AI-RTZ #926; and the long-expected Cook-to-Ternus CEO shift activated in AI-RTZ #1063.

(2) Europe’s AI Plans vs the US and China

MP TAKE: Some smart folks in Europe are reimagining AI in a “Europe 2031” thought experiment — what could happen given Europe’s more regulation- and safety-oriented, slower pace relative to companies and governments in the US and China. There’s a real sense that Europe — half a billion people — is falling behind. From my seat, the debate is absolutely useful, but Europe needs far more systemic change to be a united global player on the world AI stage.

This came on the eve of the G7, where heads of state and the senior AI company heads (OpenAI, Anthropic and others) gathered — and where, notably, President Trump said he liked Dario Amodei, even as his administration had just moved to shut down Anthropic’s latest models (our third topic). For the longest time, Europe has had a values-first regulatory plan around technology and now AI. You can point to ASML — the crown-jewel Dutch maker of the half-billion-dollar EUV lithography machines TSMC needs for the most advanced chips — but beyond that there aren’t many AI-driven champions versus the two big markets. My take: the issues go beyond AI. Europe needs to think and work as a federated, quasi-country rather than an assemblage of 27-some nations that share a currency and some unifying trading glue. It took the 13 colonies — each proud of its own beginnings and history — a great deal to become a 50-state union now creaking toward its 250th anniversary in a few weeks. Those were among the original “best laid plans,” by the writers of the Constitution and a host of folks since, each pulling their own way. Europe has all of that and more ahead of it — and AI is a good catalyst to spur some of that rethinking.

Sources, in narrative order: NYT — Why Apple’s AI upgrade for Siri won’t be available in Europe. The Guardian — A viral doomsday scenario aims to shake Europe out of its AI complacency. For longtime readers: the EU leaning back on US AI/tech in AI-RTZ #422; and Mistral, a strong AI wind from France, in AI-RTZ.

(3) The Power Struggle Behind Anthropic’s ‘Blip 2.0’

MP TAKE: Anthropic’s ‘Blip 2.0’ continues. The catalyst: two Fridays ago, the US government gave Anthropic roughly 90 minutes to pull its Mythos and Fable 5 models offline, away from anyone outside the United States. It’s deeply ironic — Anthropic has been hugely careful about safety and security since its founding, when it split off from OpenAI, with those best-laid plans as a core part of its DNA.

These newest models are 10-trillion-plus-parameter systems, versus the prior 1.5-to-2-trillion generation — super-scale, and carrying real risks from a cybersecurity perspective. And it’s not just Anthropic; it’s all the frontier models coming from OpenAI, Google and others. This was an unexpected governance shock — at a scale, speed and timing the company didn’t see coming, from the outside, and independent of its own roadmap. It’s also been a wake-up call to Europe (”should we depend on the US for technology the way we’ve depended on them for defense?”). And there’s an interesting FT piece asking whether Anthropic, with all its safety messaging, talked its way into the export ban.

Sources, in narrative order: Axios — U.S. move against Anthropic exposes AI’s missing referee. FT — Did Anthropic talk its way into an AI export ban? For longtime readers: Anthropic flagged to a screeching halt globally by the US in AI-RTZ #1117.

MP OVERALL TAKE

Stepping back to the poet Burns, each of these is an ironic example of best-laid plans. Apple thought it had a flawless strategy on design — and watched it get unraveled by AI. Europe thought it had a flawless strategy on tech and AI safety and independence — and is now debating it, right down to the nitty gritty of how Europe works vs the US and China: Apple’s Siri AI won’t be available in Europe this Fall, because of the EU’s DMA rules requiring deep interoperability. Apple is currently leveraging first-party data from its own apps (Photos, Notes, Messages, et al) and says it will work with third parties via measures that secure private data; the EU wants that data far more shareable with third parties, which Apple resists on business and real security/privacy grounds. Anthropic engineered a safety-first identity — and met an oversight system that showed up not as a process but as a 90-minute ultimatum.

MP Take: None of these were failures of laying out the best-laid plans. They were collisions between good plans and the first, second and third punch when those plans hit the real world. Plans set the direction; adaptability and resiliency usually drive what actually survives contact with a wave moving this fast — which is exactly what today’s Gadget AI, and the two questions after it, continue to underline.

Gadget AI — Roomba, the Original Proto-AI Gadget

MP Take: We’re now in the fourth year of ChatGPT, with the world investing trillions — public and private — in everything from massive AI data centers to humanoid robots, robotaxis and AI data centers in space. The Roomba story is a foretelling of this “best laid plans of mice and men — and robot scientists” dynamic. Tech happy endings take far longer than assumed, with countless unexpected twists. Worth a watch when you have the time.

Sources, in narrative order: The Verge — How Roomba started a robot revolution. For longtime readers, in narrative order: a bumpy end to a multi-decade robot story in AI-RTZ #937; Tesla’s Robotaxi & robots ride in AI-RTZ #510; and Elon Musk boosting the SpaceX/xAI pre-IPO story with ‘Terafab’ and AI data centers in space in AI-RTZ #1034.

The Verge’s “Version History” series looks at technologies that dented the world without being entirely successful — and lands on the original robot-AI company before AI was cool: iRobot, founded in 2002, maker of the Roomba vacuum. No one thought the breakthrough robot would be a vacuum cleaner — but it was. The company didn’t settle on robotic vacuums until its 12th year, after 14 different versions of what the technology might do, before landing the product that put 50 million-plus Roombas in homes (amid heavy competition from Chinese makers). The irony: it eventually went bankrupt — Amazon’s planned acquisition didn’t get the approvals — and it’s now owned by a Chinese company. Success, yet failure.

Questions

Q1 — What other widely expected AI/tech innovation didn’t deliver on time, in form and function?

Home-automation nirvana — a technology I’ve long been a keen fan of. A whole host of home-automation tech has been promised for years: lights that turn on and off on their own, smart blinds, and the rest. It’s all available, but it’s expensive, and stays expensive. After years of work the industry’s interoperability protocol, Matter, has only recently come together — the Verge has a detailed status piece on where it all stands. It’s at least 10 years behind where we thought we’d be. Best laid plans — and a theme I keep pressing from the earliest AI-RTZ days, especially when the lights are flashing green on everything from SpaceX to the latest AI gizmos.

Q2 — What other widely expected AI/tech innovation has simply taken longer than expected?

Electric air taxis — the new self-flying, helicopter-like craft. There are four-plus companies in various stages of making this a reality, but they’re tangled in suits and countersuits that keep delaying the rollout, even in the major cities that could most use these innovations. Whether it’s robotaxis on the ground or robotaxis in the air, there’s a lot beyond the technology that shapes these timelines — and changes the best-laid plans.

Full Source Reading —

For the broader context, see the canonical sources for ARD 102 — in today’s narrative order:

Theme

Event 1 — Apple’s Design Gauntlet Under New CEO John Ternus

Event 2 — Europe’s AI Plans vs the US and China

Event 3 — The Power Struggle Behind Anthropic’s ‘Blip 2.0’

Gadget AI — Roomba

Q1 + Q2 — Home automation + electric air taxis

Shorts Clips from today

Clip 1 — Apple’s AI Future: Design Challenges Ahead

Watch on YouTube Shorts

Mark Gurman reports Apple’s once-untouchable design studio has spent a decade losing altitude — no longer at the executive table, the Ive-era bench scattered to LoveFrom, OpenAI and Meta. New CEO John Ternus takes over in September with rebuilding design as job one.

MP Take: Apple spent a decade on careful post-Ive design continuity — and not rocking the boat is exactly what let it drift, just as AI rewrites what hardware design even means. Apple’s best-laid design plans get reset again, with Ternus’s first moves this Fall setting the tone.

Clip 2 — Anthropic’s AI Export Ban Surprise

Watch on YouTube Shorts

Anthropic got roughly 90 minutes to pull Fable 5 and Mythos offline after investor and partner Amazon flagged a jailbreak to Treasury Secretary Scott Bessent. The safety-first lab that publicly asked for government oversight suddenly got it — ad hoc, and triggered by a competitor’s complaint.

MP Take: An unexpected governance shock from the outside that no Anthropic roadmap could schedule. It’s less a verdict on Anthropic than a flashing sign the referee is missing — best-laid plans of a safety-first lab, undone by the absence of the very system it asked for.

Clip 3 — Home Automation: 10 Years Behind

Watch on YouTube Shorts

Home-automation nirvana has been promised for years — lights, blinds and more that just work. After years of effort the industry’s ‘Matter’ standard finally came together, but the whole space is still at least 10 years behind where we expected to be.

MP Take: Another best-laid plan that ran well ahead of reality. It’s a theme I keep pressing from the earliest AI-RTZ days — things take longer than we anticipate, exactly when the lights are flashing green on everything from SpaceX to the latest AI gizmos.

Clip 4 — Electric Air Taxis: Delays and Challenges

Watch on YouTube Shorts

Electric air taxis — self-flying, helicopter-like craft — have been the ‘next thing’ for years. Four-plus companies are racing to make them real, but they’re tangled in lawsuits and countersuits that keep delaying rollout, even in the big cities that could use them most.

MP Take: Whether it’s robotaxis on the ground or in the air, plenty beyond the technology shapes these timelines. Another best-laid plan stuck on the runway — and a reminder to temper expectations as the AI Tech Wave promises everything at once.

About AI Ramblings Daily (ARD), and AI-RTZ

Both are daily. Both are free. Both are about AI. But they’re different mediums carrying different messages.

AI-RTZ is the morning text — a deeper written take on one idea, published by at least 5 AM EST. Today: post #1125.

AI Ramblings Daily is the afternoon video + podcast — my ad hoc takes and perspective on the day’s AI issues & news flow, around 20 minutes, with short 1-2 minute clips for quick topic views. Today: episode #102.

Subscribe to either or both on michaelparekh.substack.com. They run as separate Sections you can opt into or out of.

Links used in today’s show (already embedded inline above; listed here for reference)

Theme:

Take 1 — Apple’s Design Gauntlet Under New CEO John Ternus:

Take 2 — Europe’s AI Plans vs the US and China:

Take 3 — The Power Struggle Behind Anthropic’s ‘Blip 2.0’:

Gadget AI — Roomba:

Q1 + Q2 — Home automation + electric air taxis:

Companion text:

AI Ramblings Daily on AI-RTZ is here to think through AI and reset. Together.

Today’s AI-RTZ #1125 — OpenAI and Anthropic’s tight IPO-timing gauntlet to beat SpaceX’s IPO run — SpaceX went from confidential filing to listing in a record 74 days; if OpenAI and Anthropic matched that pace we’d see IPOs before September — unlikely given the August lull — so expect those mega-AI IPOs in the September timeframe. Recommended as today’s reading post.

Tomorrow — ARD 103 on AI-RTZ 1126.

Thanks for joining us today, AI Curious Folk. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here.)

Subscribe to AI: Reset to Zero for daily AI Ramblings + Sunday Bigger Picture posts.





Want the latest?

Sign up for Michael Parekh's Newsletter below:


Subscribe Here

More like this

AI: Weekly Summary. RTZ #541
AI: Weekly Summary. RTZ #541
Michael Parekh AI: Reset to Zero
Industry concerns on Slower AI Scaling: Media reports are focused on both  OpenAI and Google potentially showing signs of slower benefits of ‘AI Scaling’,...
2 years ago
4 MIN READ
BROWSE ALL ARTICLES