- AI Fire
- Posts
- 🎭 MicroCopycat = Copylot + OpenAtlas
🎭 MicroCopycat = Copylot + OpenAtlas
AGI-B-Que SuperIntel is SuperInHell Actually! Help!?
Wendy
October 23, 2025
Is humanity ready for artificial superintelligence? Some of AI’s founding voices say no and want a global ban. But the biggest labs (I mean all the big names) stayed silent.
What are on FIRE 🔥
IN PARTNERSHIP WITH DESCRIPT
Descript makes video creation feel like editing a Google Doc. Turn your raw recordings into polished clips, product demos, or launch videos fast. Here’s how:
Edit transcripts by typing, Descript syncs your words and visuals in real time
Instantly remove filler words, background noise, and awkward pauses
Add branded B-roll, captions, and music in clicks
Collaborate with teammates, comment, and export without version chaos
Publish short clips for social in minutes, not days
Built for teams who ship fast: Used by creators and marketers at top product-led companies to move from recording to ready-to-share content - without handoffs or hassle.
AI INSIGHTS
We just got another flashpoint in the AI safety war: a new open letter backed by top thinkers is calling on governments to ban the development of artificial superintelligence (ASI).
The Future of Life Institute released the letter, urging world govs to prohibit ASI development at least until we’re 100% sure it’s controllable and society says yes.
The signers argue unchecked ASI could lead to:
Human obsolescence in the economy
Loss of civil liberties and autonomy
Complete extinction, in worst-case scenarios
Do you know who signed it? → Yoshua Bengio and Geoffrey Hinton (Godfathers of AI) + Steve Wozniak (Apple co-founder) + Leo Gao (OpenAI staffer) & more.
But notably absent leadership we noticed from the biggest AI labs like: OpenAI / Google DeepMind / Anthropic / Meta / xAI (all big names right?)
Why it matters: Just 5% want companies to build it without oversight. But there’s no clear definition of what “superintelligence” even is. And without cooperation from OpenAI, Google, or Meta, this could end up more symbolic than effective.
🎁 Today's Trivia - Vote, Learn & Win!
Get a 3-month membership at AI Fire Academy (500+ AI Workflows, AI Tutorials, AI Case Studies) just by answering the poll.
What issue makes the letter harder to enforce or act on? |
PRESENTED BY BELAY
Q4 is the perfect window to turn this year’s numbers into a clear, actionable forecast aligned with your goals. Set your business up for a stronger 2026 with BELAY’s new guide.
TODAY IN AI
AI HIGHLIGHTS
🤔 Just 48 hrs after OpenAI’s Atlas, Microsoft relaunched a nearly identical AI browser with some upgrades. Wait, did they are partners? (Microsoft invested $13B)
🍓 But then, in a head-to-head test across 4 AI browsers (all big ones), a new AI browser called Strawberry outperforms the rest by a wide margin. Here’s the result.
🌐 Is it the world builder? Tencent’s Hunyuan World open-source model now turns videos and multi-view images into detailed 3D worlds. See it in action or try it here.
💻 Google announced that its quantum chip achieved a major milestone by running on hardware 13,000x faster than top supercomputers. You can watch the video here.
👀 Microsoft just gave Copilot a face called Mico. It’s a cute & customizable AI blob that reacts to your voice (even turned into Clippy). See its cute reactions in this video.
🛒 Amazon “Help Me Decide” AI recommends products & tells you why. This feature actually helps me a lot when shopping. For me, it’s worth using, you can try it here.
💰 AI Daily Fundraising: Fal AI raised $250M, led by Kleiner Perkins and Sequoia, valuing the company at over $4B. It specializes in hosting AI models.
AI SOURCES FROM AI FIRE
NEW EMPOWERED AI TOOLS
📊 imini creates reports, presentations, docs, images in minutes
🧾 aiApply creates tailored resumes and auto-applies for jobs
🕷️ Hexomatic scrapes any contents from any website for you
🌐 CodeDesign creates & deploys your own beautiful websites
AI QUICK HITS
💪 Top AI researchers at major orgs are working 80-100 hrs/week
⚠ Warning: ChatGPT Atlas is vulnerable to attacks & against a user
📣 Reddit sued Perplexity for scraping data after licensing OpenAI & Google
🗺️ Google now turned Earth-scale data into instant insights
🤖 NVIDIA AI Day in Sydney drew 1K attendees with 600+ startups
AI CHART
We’ve seen ChatGPT ace MBA finals and hit passing scores on the USMLE. But how does it handle real medical specialization like orthopaedic surgery?
Researchers just ran ChatGPT through the OITE core test for ortho residents and it barely kept up with first-year residents:
Best year: 2021 OITE (47.3%)
Worst year: 2020 OITE (35.3%)
Overall average: Around the level of a PGY1–PGY2 (but not statistically better)
Compared to national averages:
Significantly underperformed across all PGY levels (p < 0.01)
Fell short of Jefferson PGY3–PGY5 with statistical significance
Basically, it’s expected for us cause ChatGPT is a text-based model, not designed for image-heavy questions. Once it starts a wrong explanation, it can’t “go back” and correct. But as a didactic tool, it still has real potential.
We read your emails, comments, and poll replies daily
How would you rate today’s newsletter?Your feedback helps us create the best newsletter possible |
Hit reply and say Hello – we'd love to hear from you!
Like what you're reading? Forward it to friends, and they can sign up here.
Cheers,
The AI Fire Team
Reply