The Weekly Waypoint, Issue #11
Ten weeks in. A lot has changed since Issue 1.
GPT-5 launched. Apple announced its full AI strategy. And the question shifted from "is AI real?" to "how do I keep up?"
This week: a reset. Where we are, where this is going, and what you should actually do about it, whether you're a parent, a professional, or a builder.
This Week's Big Story: The AI Landscape Just Shifted, Again
What happened: Three things converged this week that matter more than the individual headlines suggest.
First, GPT-5 officially became the default model for ChatGPT Plus subscribers. It's not just incrementally better, it reasons more reliably, remembers your preferences across sessions, and handles multi-step tasks without the constant "I lost track" failures that defined GPT-4.
Second, Apple opened Siri to third-party AI assistants in iOS 27 and restructured its entire AI division under Craig Federighi. This is Apple admitting that building AI in-house wasn't working fast enough, and opening the gates to let Anthropic, Google, and even OpenAI run on Apple devices natively.
Third, Stanford released its 2026 AI Index, over 400 pages of data showing that AI funding in Q1 2026 already surpassed all of 2023, adoption rates are accelerating in every sector, and the performance gap between open-source and closed-source models is shrinking.
Why it matters: These three developments together signal that we've moved past the "is AI real?" phase and into the "how do we actually use this?" phase. GPT-5 being the default means most people will interact with genuinely capable AI daily, not as an experiment, but as infrastructure. Apple opening up means that capability is coming to the devices people already own. And the funding data means more tools, more competition, and more options are coming fast.
The question is no longer whether AI matters. It's whether you're using it deliberately or letting it happen to you.
Quick Hits
Stanford's 2026 AI Index dropped. 400+ pages. Key takeaway: AI is being adopted faster than the internet was. If you think it's too late to get started, you're wrong, but it's getting less late.
The open-source model gap is closing. Llama 4 and Mistral's latest are within striking distance of GPT-5 on most benchmarks. For budget-conscious builders, this matters. You don't need to pay OpenAI prices for good AI anymore.
AI funding hit $97B in Q1 2026. That's not a typo. Venture money is pouring into AI infrastructure, agents, and vertical applications. More tools are coming.
Microsoft Copilot got a major update. The new Copilot can now work across all your Microsoft apps simultaneously, reading your emails, checking your calendar, drafting documents, and updating spreadsheets in a single workflow. Enterprise AI is becoming real.
The "AI skills gap" narrative is shifting. New data shows that companies investing in AI training for existing employees are outperforming companies hiring "AI specialists." Turns out, teaching your team to use AI well matters more than hiring someone with "AI" in their title.
Try This: Build Your Personal AI Stack
Stop thinking about AI as one tool. Start thinking about it as a stack, different tools for different jobs, all working together.
Here's a framework to figure out what you actually need:
Step 1: Audit your week. Write down everything you did in the last 5 work days. Highlight anything that took more than 30 minutes and follows a pattern.
Step 2: Categorize by type:
- Research & synthesis → Use Claude Opus 4.7 or GPT-5 for deep research and analysis
- Writing & editing → Use Claude for long-form, GPT-5 for quick drafts
- Repetitive tasks → Use Zapier + AI, or start building agents with the OpenAI Agents SDK
- Learning & skill-building → Use AI as a tutor, not a search engine
Step 3: Start with one. Don't try to automate everything at once. Pick the task that takes the most time and start there. Get good at one before adding another.
Step 4: Review weekly. Every Sunday, spend 10 minutes asking: "What did I do this week that AI could have done faster?" Adjust your stack.
The goal isn't to replace yourself with AI. It's to free up time for the things only you can do.
The Fun Part: AI Tried to Be a Chef This Week
A food blogger tested GPT-5's recipe generation by asking it to create "a gourmet dinner using only ingredients found in a typical American dorm room." The result? A three-course meal featuring "Instant Ramen Risotto," "Microwave-Sautéed Hot Dog Vegetables," and "Dorito-Crusted Microwave Brownie."
Honestly? The brownie recipe actually worked. The risotto was... ambitious. And the hot dog vegetables were exactly as bad as they sound.
The real takeaway: AI is creative but lacks common sense. It can combine ideas in ways humans wouldn't, but sometimes the reason humans wouldn't is because the idea is terrible. Use AI for ideation, but always apply the "does this actually make sense?" filter.
Until Next Week
Next week we're going deep on AI agents, the technology that's turning "talk to AI" into "AI does it for you." If this week's news is any indication, the agent era is arriving faster than anyone expected.
See you Monday.
Deep Dive This Week
Pro members go deeper: "The Final Vision: Your AI Future, Mapped Out", where AI is headed, what it means for your life and career, and your 12-month action plan. We covered the reset this week; the deep dive gives you the roadmap forward.