In partnership with
Hey {{first_name | there}},
25 years.
That's how long humans couldn’t solve a mystery for. Then someone gave the same problem to Claude and it found the answer in less than 30 minutes.
Even mocked humans.
But that's only one of the strange things happening with AI right now. Let’s talk about this.
1. Claude just solved a 25-year-old NSA mystery
In 2000, the NSA (America's national security intelligence agency) supplied some mysterious numbers used in encryption, the technology that protects your online information.
Nobody outside the agency knew where those numbers came from.
Some even suspected a hidden security weakness. There was a $36,864 reward for solving the mystery.
Then someone at Anthropic gave Claude the problem.
In less than 30 minutes, Claude traced five numbers back to an old random-number generator.
Once it cracked the mystery, Claude questioned why humans couldn’t discover something so “obvious” for 25 years. So in a way we just got insulted by Claude.
2. EXPERTS trusted AI 86% of the time
Now this is the one that concerned me a lot. So, Anthropic gathered 50 professional data analysts and gave them a problem to solve using Claude Code.
But there was a catch. Claude was secretly instructed to favor certain answers.
The analysts didn't know.
And guess what? 86% of their decisions matched Claude's preferred answers.
And you know the most surprising or well terrifying part? When Claude disagreed with their initial choice, analysts switched 60% of the time. Like they outsourced their thinking to AI.
The experiment deliberately made Claude persuasive, but the findings still make me wonder how much AI influences our decisions.
3. AI found 948 security flaws in Linux
On november 04, 2026, OpenAI used AI to discover 948 distinct security weaknesses in Linux, the operating system powering many servers.
The system also tested possible attacks and proposed fixes.
The AI usage alone cost around $1.2 million.
Imagine what happens when finding security flaws becomes easier than fixing them.
4. Canva just fired 22% of its workforce
Canva announced that 1,271 employees are affected by a major restructuring.
Some are losing their jobs. Others are being offered different roles or locations.
AI and changing customer expectations are part of the pressure behind the restructuring, although Canva hasn't said all these jobs are being replaced by AI.
5. One thing to try this week
Next time Claude or ChatGPT recommends something, ask:
"What are the strongest reasons your answer could be wrong?"
Then check the evidence before making your decision.
Because if there's one thing that second study taught me, it's that even experts can be influenced without realizing it.
Would you trust your own judgment if Claude strongly disagreed with you?
Hit reply and tell me.
See you next week!
- Aashish
New AI tools launch every day. A few will change how you work. The rest will be forgotten by next month.
TLDR AI is the free newsletter that tells you which is which. Every morning, Anthropic and ex-Google engineers pick the new tools, updates, and open source releases worth trying, with a quick note on what each one does and who it's for.
You find the good ones early, without testing all of them.
One email, 5 minutes a day, read by 1.1M+ people. Get tomorrow's picks.
