- Uncovering AI
- Posts
- 🔍 Google Search Doesn't Look Like Google Anymore
🔍 Google Search Doesn't Look Like Google Anymore
Apple's lawsuit against OpenAI has some brutal receipts
Sponsored by
My fellow AI explorers
I was going to open with something about the summer AI lull. There is no summer AI lull.
This week, Apple sued OpenAI, and the complaint reads less like a legal filing and more like a group chat you were not supposed to see. Meanwhile, Google quietly rewired how a quarter century of the internet works, and most people haven’t clocked it yet. Grab your coffee… this one's a lot.
In today's edition:
🍎 Apple accuses OpenAI of a coordinated trade secret heist, complete with laptop hacks and "show and tell" parts requests
🔍 Google Search is now fully AI-generated, and publisher clicks just fell off a cliff
🎁 A 5-step play to keep your content visible now that AI writes the answer page
🚀 Gemini 3.5 Pro's launch date, SK Hynix's record debut, and Meta's compute doubling
The best voice models, now across all channels
Most CX platforms do not own the voice. They orchestrate a workflow, then call a third party for speech and transcription. Every hop adds latency, cost, and another vendor to manage.
ElevenAgents is the opposite. They make the voice models the market builds on, and ElevenAgents puts full orchestration on top. Voice, transcription, text-based chat, and reasoning run in one vertically integrated pipeline, so responses come back in <400 milliseconds and sound human, not synthetic.
Plus, you keep full control. Plug in any LLM, integrate tools, webhooks, and MCP servers, and ground responses in your knowledge base. Get an agent live in minutes, then A/B test with Experiments, enforce Guardrails, and version every change.
The payoff: more human conversations, lower latency, and far less time stitching infrastructure together. You build on the models you already trust. Pricing is transparent and flat at $0.08 per minute.
Apple vs OpenAI
Apple Just Sued OpenAI, and the Complaint Is Brutal
Apple filed suit against OpenAI in a California federal court, and the headline number alone is enough to make any HR department nervous: more than 400 former Apple employees now work at OpenAI. But the number is not even the spiciest part.
Here's what the complaint alleges:
A former Apple electrical engineer, Chang Liu, kept his company laptop after leaving and later found an authentication bug that let him keep accessing Apple's internal file storage. He reportedly texted a former colleague something along the lines of "I found out I can access it, so funny."
Tang Tan, the Apple veteran who worked on the iPhone and Apple Watch and now serves as OpenAI's chief hardware officer, allegedly used Apple's internal codenames to pump job candidates for information. He asked them to bring actual hardware parts to interviews for "show and tell."
Tan is also accused of circulating an internal Apple offboarding document to teach new OpenAI hires how to dodge Apple's exit security checks.
io Products, the hardware startup OpenAI bought for $6.5 billion from former Apple design chief Jony Ive, is named as a defendant. Ive himself was not.
OpenAI's response so far: "We have no interest in other companies' trade secrets." Apple is asking the court for an injunction that would force OpenAI to stop using the disputed material entirely, which could put a real dent in its nascent hardware plans, the device Sam Altman has been teasing since November.
The timing could not be worse for OpenAI. It's reportedly prepping a confidential IPO filing with Goldman Sachs and Morgan Stanley for a possible September debut at a $730 billion valuation. A trade secrets lawsuit from Apple, of all companies, is not what the bankers ordered.
🔮 Prediction: this case drags into 2027 and never gets a clean verdict, but it does exactly what Apple wants regardless of outcome. Every AI lab just got a lot more careful about who they hire from whom and how, and the free-for-all talent raids that defined 2025 and early 2026 are about to slow down.
Reply and tell me: do you think this lawsuit is Apple protecting real IP, or Apple punishing OpenAI for becoming a hardware competitor?
Search
Google Search Doesn't Look Like Google Anymore
Three days ago, Google finished rolling out the biggest change to its core product in 27 years. Every single search query, worldwide, now returns an AI-generated answer written by Gemini 3.5 Flash as the default result. The ten blue links that trained an entire generation of internet behavior are still there, technically, but they've been pushed below the fold.
What actually changed:
Gemini 3.5 Flash reads across multiple sources and writes a prose answer in real time, with citations woven in like footnotes rather than links stacked below your query.
Google Antigravity, the agentic framework Google introduced at I/O, can now assemble that answer into custom layouts on the fly. Think interactive calculators or simulations instead of a paragraph, depending on what you asked.
There is no permanent toggle to go back to the old experience. The workaround is clicking into the "Web" filter tab after you search, and even that is not guaranteed to stay.
The fallout is already showing up in the numbers. Early data points to a 58% drop in publisher click-through rates since the rollout, and antitrust complaints are stacking up on two continents. A German court reportedly went as far as stripping Google of a liability shield tied to how it surfaces search results.
Here's the part that should land differently for you and me specifically: we run a newsletter in the AI space, which means our entire audience discovery model has, for two decades, been structured with Google rewarding good content with clicks. That deal is over. If the AI answer cites you, you exist. If it doesn't, you're invisible no matter how well you used to rank.
🔮 Prediction: SEO as we've known it is dead within 18 months, and it gets replaced by something closer to "AEO," answer engine optimization, where the unit of value is getting cited inside the answer, not ranking above it. Publishers who adapt their content to be quotable, specific, and source-worthy win. Publishers who keep writing for the old algorithm quietly disappear from the internet's attention.
Reply and tell me: has your own traffic from Google shifted in the last few days? I want real numbers from real inboxes for a future edition.
30-Second AI Play
Get Your Old Google Search Back (and See What Your Content Looks Like to AI)
Before you panic about the algorithm, do this two-minute audit:
Run a search for a topic you'd normally rank for, and read the new AI-generated answer at the top.
Check whether your site is cited inside the prose. Not ranked below it, cited inside it.
Click the "Web" filter tab under the search bar to see the classic blue-links list Google still keeps, just hidden by default.
Compare the two: does the AI answer accurately represent your point of view, or did it flatten it into something generic?
If you weren't cited, rewrite your top three pages to lead with a specific, quotable claim in the first two sentences. AI answers pull the clearest sentence, not the best-written paragraph.
💡 Pro tip: Gemini and every other AI answer engine favor content with concrete numbers, named sources, and a clear point of view over content optimized for keyword density. Write like you're being quoted, because now you literally are.
Advertise to 180k engineers and CTOs choosing what tools their companies build with.
Other Relevant AI News!
🚀 Gemini 3.5 Pro finally has a launch date of July 17, and leaked specs point to a 2 million token context window and Deep Think reasoning locked behind the $250 a month Ultra tier.
💾 SK Hynix made Wall Street history with a $28 to $29 billion Nasdaq debut, the largest ADR listing ever, because it controls 60% of the world's AI memory supply. Details here.
🏗️ Meta is planning to double its total computing power by 2027, backed by a new Samsung supply deal and a $10 billion data center in Alberta.
🧮 Two days into GPT-5.6, developers have sorted the three-tier family into roles: Terra as the value pick, Sol as the powerhouse, Luna for cheap high-volume work.
🏛️ The Federal Reserve appointed a16z's Marc Andreessen to co-lead its first AI economic impact task force, a pick that's already drawing conflict-of-interest questions.
Golden Nuggets
🍎 Apple's lawsuit against OpenAI is the AI talent war's first real day in court, and the details are ugly for OpenAI's hardware team.
🔍 Google Search is now AI-first for every query on Earth, and it just quietly rewrote the rules for how anyone gets discovered online.
🚀 Compute, memory, and context windows are all racing upward at once. Gemini 3.5 Pro, Meta, and SK Hynix are three different bets on the same underlying trend.
Would love to hear your thoughts! Send me your thoughts by replying to this email (yes, I read them all :)
Until our next AI rendezvous,
Anthony | Founder of Uncover AI