- The Upload
- Posts
- AI Gets First Dibs Now
AI Gets First Dibs Now
Shopify goes AI-first, OpenAI eyes hardware, Gemini learns screens, and China’s models are closing the gap fast.
Mark Chepelyuk
April 08, 2025
Shopify won’t hire unless AI can’t do the job. OpenAI might buy Jony Ive’s startup to design a screenless assistant. Gemini now watches your screen. And Chinese labs are pushing U.S. models harder than ever.
Today’s Upload
Shopify now requires teams to test AI before hiring
OpenAI may acquire a startup designing AI-first devices
Gemini gets real-time video and screen sharing on Android
Chinese LLMs are now competing with GPT-4 in some benchmarks
Let’s get into it. 🚀
Image source: David Fitzgerald | Sportsfile | Getty Images
🧾 Shopify Says Prove AI Can’t Do It
Before you hire anyone, AI has to get the first shot.
Key Details:
Shopify’s new policy: “AI Before People”
Teams must show a task can’t be done with AI before adding headcount
Applies to both new hires and internal role changes
First big company to make this approach official
CEO Tobi Lütke shared the policy publicly on X
Why It Matters
This is the clearest signal yet that AI isn’t just a tool, it’s a filter. Hiring is no longer about what a team needs. It’s about what AI can’t do — and that list is shrinking fast. For anyone working in ops, support, or even product, AI is now your first competition.
Image source: Deepwater
🔊 OpenAI May Acquire Jony Ive’s Startup
Altman wants to build a voice-first AI device with design roots from Apple.
Key Details:
OpenAI is in talks to acquire the AI hardware startup “Humane”
Co-founded by former Apple designer Jony Ive and ex-iPhone engineer Imran Chaudhri
Their first product, the AI Pin, launched last year with mixed reviews
OpenAI is reportedly interested in Humane’s design, talent, and form factor
Could result in a screenless, always-on, ChatGPT-native device
Why It Matters
This isn’t about another smart speaker. It’s about control over the interface. If OpenAI owns the hardware and the model, it can shape how AI is experienced from the ground up — without keyboards, screens, or apps. The phone might not get replaced, but the interaction model could.
Image source: Google
📱 Gemini Can Now Watch Your Screen
Google expands Gemini’s reach with video and screen sharing tools for Android users.
Key Details:
Gemini on Android can now view your screen and respond based on what it sees
You can record a video or start a screen share session to get help in real time
Early use cases include travel bookings, form filling, and troubleshooting
Voice support and mobile multimodality continue to expand
No iOS rollout yet
Why It Matters
This moves Gemini from a chat tool to a co-pilot. Instead of switching tabs or copying content, you can show your screen and talk. It’s faster, more intuitive, and brings us closer to agents that work across everything we do, not just what we type.
Image source: Unsplash
🇨🇳 China’s Models Are Closing the Gap
New benchmarks show Chinese LLMs approaching parity with top U.S. models.
Key Details:
Kimi, DeepSeek, and Qwen are now rivaling GPT-4 and Claude 3 on key tasks
New evals show significant progress in reasoning, math, and instruction following
Models are getting more efficient and multilingual
Many are released with open weights and available for commercial use
Part of broader state-backed AI acceleration in China
Why It Matters
The narrative that U.S. companies are way ahead is starting to crack. Chinese labs are training fast, scaling faster, and releasing models that are competitive on cost and performance. This will shape the global AI market — especially for open source, edge, and non-English use.
🕐 Quick Bits
🎨 Midjourney V7 launches
Faster, better coherence, and voice input now supported in Discord.
📦 Amazon building its own AI chips
Aims to reduce dependence on Nvidia with internal hardware by late 2025.
🧬 Colossal Biosciences revives dire wolves
The company says AI is helping reconstruct extinct DNA. Jurassic Park vibes incoming.