- Evolving AI Insights
- Posts
- 🧠 AGI still 10 years away
In partnership with
Welcome, AI enthusiasts
Former OpenAI and Tesla researcher Andrej Karpathy is pouring some cold water on the AGI hype. He says “agentic” AI systems meant to act like real assistants or employees is still years away from living up to expectations. Let’s dive in!
In today’s insights:
AGI still 10 years away
Wikipedia Traffic Slump due to AI Search
Building a New Species
Read time: 4 minutes
LATEST DEVELOPMENTS
ANDREJ KARPATHY
🧠 AGI still 10 years away
Evolving AI: Andrej Karpathy says agent-based AI is overhyped and still years from matching expectations.
Key Points:
Karpathy argues current AI models lack memory, reasoning, and multimodal depth.
He sees value in small improvements and smarter data, not big breakthroughs.
High-quality, curated training data could redefine the next wave of models.
Details:
Former OpenAI and Tesla researcher Andrej Karpathy believes the hype around agentic AI is premature. Speaking on a podcast, he said today’s models can’t act like real assistants or employees, they’re still limited to autocomplete-style tasks. Poor training data is a major issue, he says, since most internet text is low-quality junk. Karpathy expects progress over a decade, driven by smaller steps: better data, architectures, and learning methods.
Why It Matters:
The industry’s obsession with agentic AI says more about investor hype than real capability. Karpathy’s decade-long timeline feels realistic when you look at how unreliable today’s “AI interns” still are. The smarter play right now isn’t chasing autonomous agents but investing in cleaner datasets, modular AI systems, and human-in-the-loop setups that actually work. It’s less glamorous, but it’s where the real progress (and profits) are likely to show up first.
Stop Fraud Before It Starts
Are you winning but sales but losing revenue? Post-purchase fraud "digital shoplifting" slips past traditional tools and drains your revenue through false disputes, policy abuse, and wasted shipping.
Chargeflow Prevent closes that gap. Our AI-powered platform blocks fraud before fulfillment, cuts disputes by up to 90%, and saves your team wasted costs.
WIKIPEDIA
📉 Wikipedia Traffic Slump due to AI Search
Source: Getty Images
Evolving AI: Wikipedia sees fewer visitors as AI summaries and social video reshape how people find information.
Key Points:
Wikipedia traffic dropped 8% year-over-year, mostly from humans, not bots.
AI search summaries and short-form social video are changing how people seek answers.
Fewer visits could mean fewer editors and donations, threatening Wikipedia’s long-term health.
Details:
The Wikimedia Foundation confirmed an 8% decline in human visits after refining its bot detection systems. The drop highlights a shift in how people access knowledge: many now get information directly from AI search summaries or platforms like TikTok and YouTube Shorts. Foundation leader Marshall Miller noted in a blog post that while Wikipedia’s content still fuels these systems, fewer users are visiting the site itself, a trend that could slow community growth and reduce funding.
Why It Matters:
Wikipedia’s traffic slipping 8% year over year even after new bot filters is a signal that AI answers and social video are soaking up look‑ups. Pew found clicks fall from 15% to 8% when an AI summary appears. Fewer visits means fewer editors and fewer small‑dollar donors, which weakens the engine that feeds those summaries. Near term, push platforms to link and credit prominently, add UTMs to outbound citations, and test schema and structured data that AI can surface cleanly. If you publish, build native snippets for AI answer boxes and short video, then route readers to high‑intent pages.
FIGURE AI
🤖 Building a New Species
Evolving AI: Figure AI's CEO envisions robots that can reproduce and learn from each other.
Key Points:
Brett Adcock says Figure AI is creating a "new species" of humanoid robots.
He imagines self-replicating machines capable of space exploration and construction.
OpenAI and Tesla are both pushing humanoid robotics, setting up a major rivalry.
Details:
At Salesforce's Dreamforce conference, Figure AI CEO Brett Adcock told Marc Benioff the company aims to create robots that can learn, reproduce, and eventually build more of themselves, echoing sci-fi ideas of self-replicating "von Neumann" probes. OpenAI, an investor in Figure AI, shares this vision, while Tesla races to launch its own humanoid robot, Optimus, by 2026. Figure’s latest prototype, Figure 03, focuses on practical household tasks but hasn’t yet reached consumers.
Why It Matters:
Near term this race decides how work gets done. Hospitals already lean on wheeled bots like Moxi for routine runs, and Figure is lining up home and factory pilots as it readies Figure 03. Tesla signals Optimus parts orders and early production plans, pulling humanoids closer to shop floors. For operators that means voice tasking, robot friendly layouts, and data capture from human demos to train the next unit. Budgets shift from pilots to line items and rules on safety and liability get sharper. The sci fi dreams can wait, the real call today is which tasks you hand off first.
Turn AI into Your Income Engine
Ready to transform artificial intelligence from a buzzword into your personal revenue generator
HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.
Inside you'll discover:
A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential
Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background
Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve
Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.
👀 Click on the image you think is real
QUICK HITS
⚠️ Facebook’s AI Wants to Snoop on Your Photos - Even the Ones You Didn’t Plan to Share.
😮 Jensen Huang says Nvidia went from 95% market share in China to 0%.
📈 OpenAI needs new scaling laws for both its AI models and its revenue.
🎞️ The long-predicted deepfake dystopia has arrived with Sora 2.
🌐 Bid to make North Lincolnshire an AI growth zone.
⛪ People are using AI to talk to God.
📈 Trending AI Tools
🎬 Descript - Edit audio and video like text with AI*
🎨 Adobe Photoshop - now with Google’s Nano Banana build in
📱 Huxe – This mobile app creates hands-free, personalized audio briefings from your emails, calendars, and other data sources
🚀 Amazon Quick Suite – Agentic teammates for research, business insights, and automation
*partner link
What'd you think of today's edition? |
Reply