Hello and welcome back to “Let’s try to survive this crazy world,” week 43. Jumping cars, new acquisitions and investments, AI agents, and warnings about AI agents. Tech life is never boring, so I have picked the top 10 stories to discuss.
From AI-funded public good to security flaws, smarter cars, and the next MacBook debate
Tesla’s stock skyrocketed (just like Roadster?) overnight
Can we talk about Tesla’s Roadster? Even though it was last Friday, after I had already summed up my tech week brief, I saw a zillion memes of flying (sorry, jumping) Teslas and cars stuck in traffic in the air and just couldn’t pass by. Memes are funny, but the new Roadster is already called “Tesla’s best engineering model and the fastest car.” Forecasts say that there will be around 2,000 Roadster deliveries, and it will be Tesla’s lowest-volume vehicle.
The timing of this presentation is the most interesting part. SpaceX has reported a 161% revenue increase year over year, and this car could predict a potential merger of Tesla and SpaceX. Read more on the news here.
Revolut goes down worldwide
Millions of Revolut users had a frustrating morning on October 20 when a major outage disrupted the digital bank’s services. Users couldn’t log in, transfer money, withdraw cash, or pay with their cards. The issue was Google Cloud’s data center power outage, affecting businesses and public services. Every time we experience a similar issue, we prove how much of our daily lives depend on cloud providers.
Revolut reassured customers that their money and data remained safe and later confirmed that services had been restored.
Just like millions of customers, my banking experience lives in different apps, but what happens when they go down? Follow the news on Euronews.
AI is trying to save dying dating apps
Bumble and similar swipe apps are probably in the past because users are tired of useless swiping in the hope of finding a match. On the other hand, we see the rise of chatbots and AI companions that provide emotional support and even compete with dating apps. AI companions like Replika and Character.AI offer someone always available and never leaving you on read. These solutions cannot replace human interaction, but at times of loneliness, they can provide support.
So if AI has such great potential in building relationships, maybe Bumble and other apps could use it to offer better matches and help practice conversations? The long-term perspective on integrating AI is about lowering the churn rate, because when dating apps are good enough to help you find your true love, you would probably delete the app and forget. Get the full story here.
Anthropic will put 5% of its shares into the Public Benefit Fund
The CEO of Anthropic announced that the company would commit 5% of its shares to a new Public Benefit Fund, potentially worth tens of billions of dollars. The fund could support nonprofits, distribute money directly to specific groups, and help workers facing disruption from AI. As statistics show, Anthropic’s valuation steadily grows, meaning the stake will increase too. The company says an additional 2% could be added for every 50% increase in the company’s value.
The announcements come as AI companies prepare for massive IPOs while warning that their technology could have “catastrophic or existential risks to humanity,” disrupt jobs, and reshape the economy. Is this 5% shared to shut our eyes and make us numb to the emerging danger? Check the article to find out more about the news and what Elon Musk has announced.
Google’s Gemini is now in our cars
Artificial intelligence is part of our lives, and from now on, we have a chance to get personal intelligence while driving. Google is bringing Gemini’s Personal Intelligence to Android Auto, making a car assistant a little less robotic and a lot more personal. Instead of simply following commands, the car assistant now offers tailored help using information from connected Google apps. Ask for a restaurant recommendation, and it could suggest a place based on your preferences.
The idea is to make driving more convenient while keeping your hands on the wheel and eyes on the road. But the more useful your assistant becomes, the more personal information it needs access to.
Your car already knows where you’re going. Now your AI wants to know why. See the full story here.
AI found a PHP security flaw: websites are going dark
The morning of 22 October started with government agencies shutting down their websites after critical security vulnerabilities were found in websites built on PHP. The flaws could allow attackers to execute malicious code remotely, potentially compromising servers and sensitive data. Security researchers used AI to identify the vulnerabilities, and experts warned that hackers could also use AI to exploit them.
PHP developers are now working on security fixes, but 70% of the world’s websites are now being updated, so they may work with disruptions.
This time, AI was on the good side, but is it permanent? Read more about the news here.
The new Mercury 3 AI model claims to outperform GPT-6 Luna and Claude Haiku 5.5
Inception, a California-based AI product company, made bold announcements that its new model runs eight times faster than its popular rivals, generating 1000 tokens per second. Its secret is a different approach called diffusion: instead of generating text one token at a time, Mercury 3 predicts and refines multiple tokens in parallel. It also supports a one-million-token context window, making it best for coding, real-time translation, and AI agents that need to respond quickly.
These numbers are, of course, impressive, but the thing is that speed and quality figures come from Inception’s own benchmarks, and we are still waiting for results from independent testing. More details here.
Nvidia presented its cute tiny robots
Forget humanoid robots for a moment. MicroDucks, presented on Nvidia’s GTC Berlin stage, demonstrate how small robots can coordinate their movements and work together. Each robot follows simple rules and responds to its surroundings, allowing the group to perform tasks as a coordinated system.
Jensen Huang, who led the presentation, gave the ducks a few commands that looked really cute, making the audience laugh and cheer them. These robots could be useful for proving physical AI concepts and prototyping, where inexpensive robots replace large machines. Catch the full story here.
Microsoft presented new security features to protect against rogue AI agents
Increasing warnings about the AI threat forced companies to act, and Microsoft reacted with Guardian Agents, a new feature in Project Perception that monitors AI agents. This new system is meant to address the hacking models of Anthropic and OpenAI. While the threat is mostly outside, Microsoft CEO Satya Nadella commented that businesses tend to trust advanced AI models blindly, and the threat can be equally real from inside. The concern is that agents can access files, send emails, run code, and interact with company systems, sometimes making harmful decisions even without a malicious attacker.
We’ve gone from asking AI to write emails to worrying it might send the wrong ones to everyone in the company. And this is not the worst scenario. More about the news on TechCrunch.
Apple’s new MacBook with touchscreen
The new product announcement, scheduled for October 28, has some speculation, and one rumor is the most exciting. Analysts predict we will have a touchscreen on a MacBook, a feature that has kept such interaction away from laptops, insisting that laptops are better suited to keyboards and trackpads.
MacBook sales account for only 10% of Apple’s revenue, but maybe, just maybe, a new touchscreen feature will change the picture. Next week we will talk about this event. For now, read the news for more information.
Stay tuned with us. Join Intech’s Express Newsletter and get a weekly digest of the tech industry. We don’t send promotional and marketing updates. We send ONLY NEWS and only ONCE PER WEEK :)