How Caffeine alters the human brain
A summary of things I read this week
I read this really interesting article about how caffeine alters the human brain’s electrical braking system. Researchers found that caffeine delivered via gum alters the brain’s “electrical braking system,” which is linked to inhibitory control, reaction time, and cognitive processing.
The study suggested that caffeine may temporarily improve alertness and mental performance by altering how the brain regulates attention and response inhibition, offering another example of how quickly absorbed caffeine sources can directly influence neural activity. So, is it good news for coffee lovers? Maybe, but I have also found that too much coffee creates some sort of anxiety, too, so balance it out.
If you are not following the OpenAI and Anthropic announcements about Forward Deployed Engineers (FDE), this is for you. OpenAI recently announced the launch of OpenAI Deployment Company
They also acquired a company of 150 FDEs. So what’s going on? With all these Frontier AI models solving complex problems, there’s an increased demand to deploy them in large enterprises for use cases specifically tailored to them. This is where FDEs come in. FDEs work directly with large enterprises, understand their problems, and build solutions for them. FDEs are not consultants; they don’t just suggest things, they go and build things. I think this is especially useful given that all these large enterprises have an extreme amount of proprietary data and flows. So both Anthropic and OpenAI want to make sure their models get those customers.
Robotics is a field that will change rapidly over the next 3-5 years as foundation models become increasingly powerful. Unitree recently released A Manned Transformable Mecha
This is wild. It’s the first production-ready mecha. Do watch the video shared in this post. I believe personal robotics will be a crazy big industry in the next 5 years. We already see work-specific robots like Matic or Lume from Syncere. But I think something like Optimus from Tesla, and more generic robots, will start reaching consumers. The future is gonna be really wild, where you are asleep, and your personal AI-enabled robot is finishing your chores.
Another example is the Figure Helix-02 robot, which can now sort packages at the human level. They previously tried a 1-hour shift and have now done 8 hours. Industrial humanoid robots are also something I am optimistic about.
While it all sounds good, it also increases the inherent security risk of someone compromising those machines and doing something unexpected. So with all these cool things, there come some risks too. We will all see how this plays out, but I am excited. What do you think about this? Share below in comments.
So there are a lot of discussions about how AI is actually making humans stupid. Humans are trusting AI results without actually understanding how they came about. This actually presents an opportunity for how learning systems of the future will be. One example is Oboe Labs
They are building a learning system powered by AI to help you learn new concepts. It will act like an interactive teacher, keeping you engaged. You can treat it like your personal teacher and ask questions anytime.
I think this is a great concept. We as humans should keep making it easier for anyone to learn more, even in this age where we can just delete anything and everything to AI. Understanding things from first principles will always have value.
Interesting Posts to read
https://datacenters.mts.now/follow-the-risk
Did you read anything interesting last week that is worth sharing? Share it in the comments.