- Outskill with AI
- Posts
- too dependent on ai already?
too dependent on ai already?
Yup, it’s a bit crazy
In partnership with
Today’s edition is a reality check for you.
So let’s start with a poll:
How do you use AI? |
Anthropic's Economic Index report just got released.
AND
Across all users of Anthropic / Claude, 4/10 use AI to do their work.
To be specific, this number jumped from 27% to 39% in eight months.
To be more specific, these are people giving Claude complete tasks and walking away. No back-and-forth. No iteration. No learning from the process.
The opposite is "augmentation" where humans collaborate with AI. They ask follow-up questions. They build on AI insights. They iterate together to reach better solutions.
And here's the part that should terrify everyone: This isn't happening equally across the world.
Yup, it’s a bit crazy. Before we dive further, let’s cover what’s happened this week.
AI that’s trending:
1. Google TV Adds Gemini AI To Make Your TV More Conversational.
Google integrated Gemini into TVs starting with TCL's QM9K model, letting you ask vague questions like "find something for different friend groups" or "that hospital drama everyone's talking about." The AI works across multiple tabs, has far-field microphones, and plans to expand to Google TV Streamer and other brands before year-end.
Read more
2. Nvidia to Invest $100 Billion in OpenAI.
Nvidia announced plans to invest up to $100 billion in OpenAI while deploying 10 gigawatts of AI data centers with millions of GPUs. The partnership starts in late 2026 with Nvidia's Vera Rubin platform, representing the largest corporate tech investment in history. OpenAI pays cash for chips while Nvidia gets equity stakes.
3. OpenAI CEO Sam Altman Shares Which Jobs AI Could Replace.
Sam Altman told Tucker Carlson that customer service jobs will be "totally, totally gone" as AI handles phone and computer support better than humans. He said programming jobs will evolve rather than disappear, while nursing and other human-connection roles remain safe. Altman admits losing sleep over AI's real-world impact on employment.
AI Tools That Made Me Question My Life Choices This Week
Perplexity rolled out an alpha email assistant that automatically scans, labels, and manages your inbox. It auto-drafts replies you can edit before sending, schedules meetings with customizable buffer times, and can be CC'd on emails to handle entire conversations on your behalf. The assistant handles both organizing and responding autonomously. Bye to poke?
Chrome launched its biggest AI upgrade ever, adding Gemini integration that works across multiple tabs, recalls past websites from memory, and will soon handle tasks like booking appointments autonomously. Also includes AI-powered scam detection, one-click password changes, and contextual search suggestions right from the address bar.
Ray3 is the first AI video model that can think, evaluate its own outputs, and refine results like a creative partner. It generates true HDR video in professional ACES format and includes Draft Mode for 20x faster ideation. Adobe integrated it into Firefly as the first external partner, with major agencies like Dentsu Digital adopting it.
Back to the report, and about that data.
Picture two software developers. Both have access to the same AI tools.
After six months, Developer A can prompt AI to solve problems but has no idea how the solutions work. Developer B can collaborate with AI to solve problems they never could have tackled alone.
Here's what Anthropic's data reveals: Entire countries are becoming Developer A.
The Geographic Divide Nobody's Talking About
The report shows something stark. Countries with higher AI usage per capita are more likely to collaborate with AI. Countries with lower usage just delegate tasks to it.
But here's what's really revealing:
India's pattern: Over half of all AI usage focuses on coding tasks, compared to roughly a third globally. The typical interaction: "Build me this function" → AI delivers → conversation ends. India's overrepresented requests? Almost exclusively software development.
Singapore's pattern: AI usage is spread across education, scientific research, and business operations. The typical interaction: AI suggests approach → human asks follow-up questions → iterates on solution → builds on insights.
Brazil's twist: Heavy usage for translation and legal services - leveraging local specialization in AI-powered judicial systems.
Vietnam's approach: Software development AND education, showing they're building AI collaboration skills beyond just technical tasks.
The US reality: Overrepresented requests include household management, job searching, medical guidance: using AI across diverse life domains, not just work tasks.
Both Singapore and India got their immediate tasks done.
But only one population got systematically smarter in the process.
Scale this across millions of people over months, and you see the problem.
Some countries are creating populations that get more intelligent by working with AI. Others are creating populations that get more dependent on AI.
The Enterprise Warning
When companies use AI through APIs / meaning they've built it into their actual business processes: 77% of usage is pure task delegation.
The specific breakdown:
But here's what's shocking: The most expensive AI tasks get used the most. Companies don't care about AI costs when the value is clear. A 10% price drop only increases usage by 3%.
Translation: Capability matters infinitely more than cost.
Think about what this means. Businesses have figured out that it's operationally easier to have AI complete entire workflows than to train humans to collaborate with AI effectively. They're systematically choosing to automate their highest-value work first.
But here's the trap: What happens when your entire business runs on AI systems you don't understand?
Last month, a mid-size marketing agency we work with had their AI writing tool go down for a day days. Their copywriters, who had spent two years delegating all writing tasks to AI, couldn't produce client work at the quality their clients expected. They had outsourced not just the work, but the skill itself.
This is happening at national scale.
The Context Catastrophe
The report reveals that complex AI tasks require 4x more background information than simple ones.
But there's a cruel pattern in the data: The countries already using AI sophisticatedly are the same ones with organized information infrastructure to provide rich context. The countries using AI basically can only provide basic prompts.
What this looks like in practice:
The first query gets sophisticated strategic analysis. The second gets generic business tips.
The enterprise data proves this pattern: API customers using Claude for complex tasks provide dramatically longer inputs. Tasks typical of computer and mathematical occupations cost 50% more than basic tasks due to context requirements, yet they dominate usage because the economic value justifies the expense.
Countries and companies that can organize and feed rich context to AI get exponentially better results. Those that can't provide good context get basic responses.
But here's the cruel irony: The same places defaulting to "just do my work" AI usage also tend to have poor information systems.
So they're training their people to depend on AI while simultaneously being unable to provide AI with the context it needs for sophisticated work.
They're creating AI-dependent humans who can't even access good AI capabilities.
The Wake-Up Call
Ask yourself: whether AI is making you more intelligent or just more efficient.
Because there's a massive difference between the two. And only one of them has a future.
Until next time,
Vaibhav 🤝
If you read till here, you might find this interesting
#AD 1
$10k free to test catalog ads on TV
When the leader in catalog ads partners with the leader in CTV, something special happens.
Catalog ads on streaming TV with a free $10k to test it.
Turn your product feed into high-performing video ads on the biggest screen in your customers' homes.
CTV + catalog ads = your next massive growth opportunity.
#AD 2
Startups who switch to Intercom can save up to $12,000/year
Startups who read beehiiv can receive a 90% discount on Intercom's AI-first customer service platform, plus Fin—the #1 AI agent for customer service—free for a full year.
That's like having a full-time human support agent at no cost.
What’s included?
6 Advanced Seats
Fin Copilot for free
300 Fin Resolutions per month
Who’s eligible?
Intercom’s program is for high-growth, high-potential companies that are:
Up to series A (including A)
Currently not an Intercom customer
Up to 15 employees
