Apple’s new research paper dismantles the myth of AI reasoning, revealing that models from OpenAI, Anthropic, and Google collapse under complex tasks. Released ahead of WWDC 2025, the findings challenge billion-dollar AGI claims and expose the industry’s most persuasive illusion.
At WWDC 2025, Apple introduced its biggest design overhaul in a decade with Liquid Glass, but failed to deliver major AI breakthroughs. Combined with a critical research paper and growing industry pressure, the event left many questioning Apple’s pace in the AI race.
DocketAI, founded by former ZoomInfo and FullContact execs, is transforming B2B revenue operations with AI agents that automate sales engineering tasks, reduce overhead, and improve win rates—driving enterprise adoption and investor confidence.
WWDC 2025: A Polished Interface, But Apple’s AI Struggles Remain
At WWDC 2025, Apple introduced its biggest design overhaul in a decade with Liquid Glass, but failed to deliver major AI breakthroughs. Combined with a critical research paper and growing industry pressure, the event left many questioning Apple’s pace in the AI race.
When Tim Cook opened WWDC at Apple Park, he acknowledged the challenges facing the company with a clear and direct message:
“This year, we face headwinds in AI, regulatory scrutiny, and tough questions from our developer community. We will continue to invest in technologies that shape the future—artificial intelligence, spatial computing and privacy remain our top priorities.”
With competitors moving quickly on AI, Apple needed an event that delivered both new ideas and practical results. Cook’s remarks set the tone for a keynote that aimed to balance long-term vision with immediate improvements across Apple’s ecosystem.
Liquid Glass: A Fresh Look Across Devices
The main announcement was Liquid Glass, Apple’s biggest design overhaul since iOS 7. Craig Federighi explained how translucent menus, adaptive toolbars and floating icons bring a sense of depth and clarity to every screen. He called it
“the most expressive and delightful interface we’ve ever built.”
Expressive. Delightful. But still instantly familiar.
iOS 26 and iPadOS 26: A Clear theme, floating windows and streamlined navigation.
macOS Tahoe 26: Live Activities in the menu bar and seamless Phone integration.
visionOS 26: Lifelike “Personas” avatars and anchored virtual widgets.
watchOS 26 and tvOS 26: Frosted-glass effects and updated controls for fitness and media.
By unifying these elements, Apple hopes everyday tasks—from reading notifications to organizing apps—feel smoother and more connected. Watch the entire WWDC25 keynote presentation below:
Practical AI Improvements
Instead of a dramatic Siri moment, Apple focused on real-world Apple Intelligence enhancements. FaceTime, Messages and phone calls now offer live translation to “remove language barriers for good.” Visual Intelligence lets you tap any on-screen content—photos, screenshots or articles—and ask questions powered by on-device models.
Craig Federighi noted, “Some of our Apple Intelligence features needed more time to meet our exacting standards. We’ll bring those later this year and into next.”
This approach emphasizes privacy and performance, even if it meant delaying certain features.
“The Illusion of Thinking” Paper
Ahead of WWDC, Apple released The Illusion of Thinking, a research paper examining whether today’s top AI models truly reason or simply pattern-match. The study tested models like OpenAI’s o3, Anthropic’s Claude 3.7 Sonnet, DeepSeek R1 and Google’s Gemini on logic puzzles of increasing difficulty. It found that:
The results revealed a clear pattern. On low-complexity puzzles, standard language models often outperformed the more elaborate reasoning models. As the problems grew moderately harder, reasoning models began to show an edge, demonstrating some ability to handle increased complexity. However, once tasks reached high complexity, every model suffered what the authors term a “complete accuracy collapse,” frequently giving up before even attempting to solve the toughest puzzles.
“Current evaluations focus primarily on established mathematical and coding benchmarks, emphasizing final answer accuracy. However, this paradigm often suffers from data contamination and fails to provide insights into the structure and quality of reasoning traces.”
By calling attention to the reasoning process itself, Apple argues for more detailed measures of AI performance.
Industry Reactions
Nathaniel Whittemore’s AI Daily Brief podcast examined Apple’s WWDC strategy alongside The Illusion of Thinking paper. Whittemore observed that despite launching “Apple Intelligence” last year, WWDC 2025 was “notably devoid of significant AI announcements,” and he called Siri “an absolute disgrace” for its continued limitations.
Industry voices echoed his critique.
Linus Ekenstam labeled this “potentially the most boring WWDC ever” and dismissed the new Glass UI as a “UX nightmare.”
My thoughts on today’s Apple Event.
Apple has clearly missed the mark for far too many times now. I feel today was yet another one of these occurrences. Sadly
Apple is trying too hard to do too much. There is too much fat, need to trim down all that fat. Back to basics. Apple…
Portfolio manager Andrew Choi warned that Apple’s lack of AI progress represents “an existential risk.” By contrast, Bloomberg’s Mark Gurman praised the event’s “cohesive story” and productivity features—but still noted the “startling” absence of AI breakthroughs.
Excellent WWDC: Cohesive story, deep integration and continuity across the devices, zero false promises, impressive new UI and significant new productivity features on the Mac and iPad. But the lack of any real new AI features, despite that being my expectation, is startling.
Investor Scrutiny Intensifies as Apple’s WWDC Faces AI Expectations
Apple’s WWDC is under harsh investor and market analyst spotlights this year, as questions mount over whether the company is falling behind in the artificial intelligence race. While rivals like Microsoft and Alphabet have captured headlines with bold AI innovations, Apple’s developer conference was marked by a cautious, privacy-focused approach and only incremental updates to its Apple Intelligence suite. The absence of breakthrough features—most notably, the delayed overhaul of Siri—has left many observers underwhelmed, a sentiment reflected in Apple’s 19% share price decline this year, making it the biggest drag on the Nasdaq 100.
Still, not all industry voices are sounding the alarm. Bob O’Donnell of TECHnalysis Research points out that it’s far from clear whether consumers are basing hardware purchases on AI features alone, and Apple’s strategy of seamless integration may still appeal to its loyal base.
With growth slowing and competitive pressure intensifying, however, Apple faces mounting urgency to demonstrate its relevance and leadership in the AI era.
Building Developer Confidence
Cook closed the keynote by reminding developers that betas are available immediately, with public releases next month. He stressed Apple’s commitment:
“Our focus remains on building seamless, reliable experiences that respect your privacy. The best is yet to come.”
While some praised Apple’s careful approach, others worry that without bolder AI moves, the company may fall further behind rivals who are integrating AI more aggressively.
Looking Ahead
WWDC 2025 reinforced Apple’s design strengths through Liquid Glass and delivered useful AI updates—live translation, Visual Intelligence and new on-device tools. The Illusion of Thinking paper and podcast critiques highlight Apple’s cautious stance, even as competitors push forward. Now, Apple’s task is to turn thoughtful research and small-scale features into meaningful advances that keep it competitive.
Sign up for Cyber News Centre
Where cybersecurity meets innovation, the CNC team delivers AI and tech breakthroughs for our digital future. We analyze incidents, data, and insights to keep you informed, secure, and ahead.
Apple’s new research paper dismantles the myth of AI reasoning, revealing that models from OpenAI, Anthropic, and Google collapse under complex tasks. Released ahead of WWDC 2025, the findings challenge billion-dollar AGI claims and expose the industry’s most persuasive illusion.
Palantir’s AI tools are reshaping global power—from military intelligence to IRS audits. As CEO Alex Karp defends the company’s role in modern governance, critics warn of creeping surveillance. Is Palantir protecting democracy—or quietly redefining it?
NVIDIA CEO Jensen Huang has turned geopolitical risk into strategic advantage, driving $44.1B in quarterly revenue and surging AI chip demand. Despite export curbs and rising China competition, Nvidia’s Gulf alliances and Blackwell GPU rollout cement its global dominance.
Jensen Huang spearheaded Trump’s assertive AI strategy, driving Nvidia’s profits up 69% despite intense US-China tensions. Together with Elon Musk, Huang orchestrated landmark Gulf deals, embedding American tech globally, boosting Silicon Valley dominance, and sidelining China's AI ambitions.
Where cybersecurity meets innovation, the CNC team delivers AI and tech breakthroughs for our digital future. We analyze incidents, data, and insights to keep you informed, secure, and ahead. Sign up for free!