- WKND AI
- Posts
- Your Year With ChatGPT
Your Year With ChatGPT
+Waymo Causes San Fran Traffic Jam
Hello WKND AI Warriors!
OpenAI launches “Your Year with ChatGPT,” a personalized year-end recap showing how you used the chatbot in 2025.
Also, Anthropic open-sources Bloom, a new framework to automatically evaluate dangerous and misaligned AI behaviors across frontier models.
Plus, Nvidia gains access to Groq’s ultra-fast inference tech as China unveils a light-powered AI chip that outperforms GPUs by 100x on efficiency
Oh yeah, and San Francisco traffic grinds to a halt after a power outage freezes Waymo robotaxis across multiple neighborhoods.
Today’s newsletter includes:
🎓 AI COURSE OF THE WEEK
📰 AI NEWS RECAP
🤿 AI DEEP DIVE
📝 AI PROMPT OF THE WEEK
🎨 AI IMAGE OF THE WEEK
🎓 AI COURSE OF THE WEEK
Tired of feeling behind on AI but don’t have time for a 20‑hour course you’ll never finish?
This is a weekend‑sized bootcamp built for 9‑to‑5ers who want 5–10 hours of their week back without becoming “AI people.”
You get a 60‑min crash course, a 4‑hour deep dive, and plug‑and‑play workflows you can use Monday morning.
👉 Register here and have your first AI workflow live before next week.
📰 AI NEWS RECAP
Your Year With ChatGPT
ChatGPT just gave everyone a Year End Wrapped.
Like Spotify, but for how you actually used AI.
Not vanity stats. Patterns.
It shows what you asked most.
Where you relied on it.
How AI fit into your work and learning this year.
Why this matters
Your prompts are a mirror.
They reveal where you struggle.
Where you save time.
Where AI genuinely changed how you work.
If you mostly used it to rewrite emails or summarize docs, that says one thing.
If you leaned on it for planning, reasoning, or learning, that says another.
This is not about how much you used AI.
It is about how you used it.
The people who reflect on that will use AI better next year.
Anthropic released Bloom, an open source framework that auto-generates scenarios to test specific AI behaviors and score how often they occur. It has already been used to measure traits like sycophancy, sabotage, and self-preferential bias across multiple frontier models.
Manus now lets you precisely edit Nano Banana Pro–generated slides by clicking to change text or local visuals without regenerating the whole image. This combines fast AI slide creation with the fine-grained control of traditional design tools.
A San Francisco power outage froze clusters of Waymo robotaxis in place, backing up traffic and blocking lanes in several neighborhoods. The incident highlighted how autonomous vehicles can fail when city infrastructure and connectivity go down.
Groq and Nvidia signed a non-exclusive deal letting Nvidia license Groq’s high-speed inference technology while hiring key Groq executives. Groq will keep running GroqCloud independently as both companies pitch the agreement as a way to scale low-latency AI inference.
OpenAI described new safeguards to harden its Atlas AI browser against prompt injection after tests showed malicious pages could hijack tasks. The company added adversarial training, automatic attack detection, and stricter system controls, while admitting this attack class can’t be fully eliminated.
Chinese researchers unveiled LightGen, a photonic AI chip that uses light instead of electricity to run models. In tests, it delivered over 100x better speed and energy efficiency than Nvidia’s A100 on several generative image tasks.
🤿 AI DEEP DIVE
Are you getting the most out of ChatGPT?
Have you tried customizing it specifically for you?
If not, watch this.
📝 AI PROMPT OF THE WEEK
Copy and paste this into your favorite chatbot.
USE CASE: Personalized sales email.
PERSONA: SDR coach.
CONTEXT: I will paste a prospect profile and my offer.
TASK: Draft three short first-touch email variants with a relevant opener and CTA.
FORMAT: Three variants at 80 to 110 words each with a one line subject; include a clear CTA in each.
TONE: Warm, credible, value led.Master prompting with the PCTFT framework.
It turns generic AI outputs into laser focused insights.
Why it works?
Persona: Assigns a specific role to the AI.
Context: Provides background information.
Task: Defines the specific action or response needed.
Format: Specifies the structure of the output.
Tone: Indicates the desired style or mood.
🎨 AI IMAGE OF THE WEEK
Give AI image creation a try for yourself.
Copy the text below into Google’s AI image creator!
Ferrari 250 GT Pininfarina Coupe, 1959, blue and silver gradient, light brown leather seats, in front of a villa with columnsHow'd you like this newsletter?Love it or hate it? Let us know why! |
How can you help?
Refer my newsletter to help others learn AI.
Missed last week’s edition?
