AI workflows are quickly becoming part of everyday life, not just business tools or tech experiments. In this Week 13 discussion, we catch up on major AI developments since the holidays and focus on where AI actually helps, from learning faster to solving real-world problems.
Here’s what stood out.
1. Google is winning distribution (and that matters more than model quality)
Samsung is shipping a massive amount of “AI phones” this year, and they’re powered by Google’s Gemini.
That’s not a small detail.
When AI comes pre-installed, the average person doesn’t shop for a model. They use what’s already there. That’s how platforms win. It’s how Google wins.
The takeaway:
If you’re building workflows for a team or business, it may not matter what you personally prefer. It matters what your ecosystem defaults to.
2. CES is shifting from EV hype to autonomy hype
A big theme at CES is that auto companies are shifting focus away from EVs and leaning harder into autonomous driving.
The punchline was basically:
People don’t want to charge batteries. They want to stop driving.
Whether you love that or hate it, it’s a signal of where product investment is going.
3. Claude can help with things like YouTube thumbnail systems
There’s a tutorial floating around for using Claude to design YouTube thumbnails.
The interesting part isn’t “Claude can make an image.” The interesting part is building a repeatable system:
- apply brand guidelines
- match high-performing patterns
- standardize thumbnail style
- reduce decision fatigue
That’s the real win. Less creative chaos. More consistency.
If we want to test this, we can create a “thumbnail workflow” that outputs:
headline options, layout suggestions, brand-safe colors/fonts, and style variations.
4. The breadmaker example is the best “AI use case” in the episode
Someone got a breadmaker for Christmas and used AI as a diagnostic partner.
Not “give me a recipe.”
More like:
- here’s what I see
- here’s what I smell
- here’s what it feels like
- help me adjust one variable at a time
That’s the right way to use AI.
AI as a training partner. Not a crutch.
Same idea applies to real skills:
cooking
DIY repairs
accounting
even cutting hair (yes, that came up)
The win is not “AI did it for me.”
The win is “AI helped me learn faster.”
5. Agents + code: branches, safety, and using AI without wrecking the repo
We touched on Codex and GitHub-style agent workflows.
If we ever go down that route (or even just keep using Copilot), one rule matters:
Don’t let AI write directly to main.
Use branches. Test. Then merge.
That one habit prevents a lot of pain.
6. Bigger context windows change what tools are best
There was a solid reminder here: different AI tools have different context windows (how much they can “hold” at once).
This matters if you’re trying to do things like:
- upload a whole book
- load long transcripts
- work inside large documentation sets
- analyze large business context
Gemini and Claude can handle much larger inputs than many ChatGPT workflows, depending on model/version.
But the key takeaway was even simpler:
Don’t assume. Ask the model what its context window is.
Because search results and charts are often wrong or outdated.
7. Reading workflows: AI can help you retain, not just summarize
One workflow shared was: upload an ebook, have AI explain chapter by chapter, relate concepts to earlier sections, then quiz you.
That last part is the important part.
Quizzing forces recall. Recall creates retention.
So instead of always asking AI questions, flip it:
Ask AI to ask you questions.
That’s where learning actually sticks.
8. AI shopping agents are here (and people are willing to trust them)
Price.com has a “buy with AI” feature that can manage shopping end-to-end.
The key stat mentioned: a meaningful chunk of consumers are open to letting AI handle routine shopping.
The takeaway: this is going mainstream faster than people think. Agents won’t just “answer questions.” They’ll complete transactions.
9. AI content farms are already flooding platforms
There was a blunt segment on AI-generated YouTube Shorts and “brain rot” content.
Whether the percentages are exact or not, the trend is obvious:
Cheap AI content is flooding feeds.
This probably won’t last forever in its current form, but it’s here now.
Same thing is happening in music too. Some AI songs are hard to detect because they’re “too perfect.”
Short-term: noise increases.
Long-term: trust and authenticity become more valuable.
10. Voice is becoming the default interface
Voice AI is growing fast for one reason:
Speaking is faster than typing and accuracy is now extremely high.
This matters for personal workflows. If you’ve spent decades typing (like I have), voice feels unnatural at first.
But it’s worth retraining for specific use cases.
Example: weekly check-ins or reflections
Talking into AI may become the fastest way to think out loud, organize, and turn chaos into a plan.
11. The furnace story shows where this is going
One of the most practical examples:
Someone’s furnace died.
They uploaded a photo of an error code and their HVAC setup.
AI spotted a hidden safety switch and diagnosed the actual issue.
Suggested a simple fix. Heat came back in 15 minutes.
This is where AI becomes a real “everyday problem solver,” not a business toy.
We used to learn these skills from parents.
Then we used Google and YouTube.
Now we’re moving into: “show AI the situation and get step-by-step guidance.”
And eventually this will become:
- AI draws the fix
- AI generates a video tutorial
- AI tells you what to buy and how to do it
12. The bigger shift: technical skills get automated, human skills become the differentiator
There was a note about AI models scoring extremely high on finance exams, and how fast that progress has been.
The point isn’t “finance is dead.”
The point is: technical knowledge is getting commoditized.
So the value moves toward:
- judgment
- relationship management
- decision-making
- leadership
- context and nuance
That’s the real competitive moat.
What we’re taking from Week 13
Three clean takeaways:
- AI is becoming normal infrastructure (phones, voice, shopping, home troubleshooting).
- The best use of AI is skill-building (not laziness).
- Tool choice matters because context windows and ecosystems change what’s possible.
That’s the episode.
