Thoughts

Thoughts

Alessandro Perilli
The inequality of defeating death
Should we unlock next-gen medicine (e.g. body repair) without ALSO unlocking infinite abundance, we might end up in a world where the human body becomes a prison for poor people. Or one even more unequal, where only rich people live forever.

Alessandro Perilli
Turning AI censorship into a mundane task
A large language model is, at its core, an aid to express ourselves. It's troubling how certain AI providers are censoring words in the Western world even when those words are simply the names of their competitors. And it's dangerous that nobody is complaining.

Alessandro Perilli
Do we want scientists to invent AI or God?
We want AI to be flawless in bias, judgement, and morals (even if they have been trained on an immense pile of flawed human output). But isn't that a characteristic of "God" in multiple religions? Are we asking scientists to invent the digital mirror of our God?

Alessandro Perilli
Humans are the only species that won't take no for an answer
After a lifetime of reflection, I've come to the conclusion that the defining trait of the human species is not our distinctive cognitive abilities per se. It's not taking "no" for an answer. We use our distinctive cognitive abilities for that, to brush aside every no.

Alessandro Perilli
AI vs Reality, Truth, and Faithfulness
We have to be very careful when we talk about LLM hallucinations. Hallucinating compared to what reality? For many humans, Reality = Truth = "Faithfulness to something I experienced or learned, directly or via others." So what's the Truth we can measure AI hallucinations against?

Alessandro Perilli
The Future of User Interfaces
Our hands will replace mouse and trackpad, becoming the fastest and most intuitive user interface we ever had. Our smart watches will become the only peripheral we'll need to make it happen. The rest, the eye-tracking technology, will come as part of our computers' webcam.

Alessandro Perilli
People as Blueprints
People are living and breathing blueprints, templates, patterns that we can imitate and make our own. If we are lucky enough, we encounter a blueprint that changes our life. Just 100y ago, books were almost the only way to make those encounters. Today, just open your browser.

Alessandro Perilli
What does destiny look like in pictures?
I saved the article featuring this picture for 15 days, hoping to find the right words to describe how I feel looking at it. There's something about the image that is unsettling. A world that fades away. I found myself wondering: what does destiny look like in pictures?

Alessandro Perilli
AI is for silly questions
My most enjoyable use of generative AI: I have a lot of ideas and I always wanted to know if there's a philosophical or a scientific theory that resembles them. Trying to do this with Google has always been close to impossible. With GTP-4, it's a playground for the curious mind.

Alessandro Perilli
How do we measure comprehension now?
The challenge of every educator will soon be the same of every employer: how can I be sure that this person has understood the topic well enough to use the knowledge in new situations?

Alessandro Perilli
A safer kind of rebel genius, please
A reason why people might prefer AI artists in the future. A rebel genius with a switch on/off button sounds a lot safer.

Alessandro Perilli
Resurrecting the Dead
We should expect that, going forward, every famous singer or actor ever existed will be resurrected with AI and turned into a synthetic celebrity. In the same way, the ones that we have today will live forever. It's a strange way to defeat death, don't you think?

Alessandro Perilli
Frank Sinatra sings Gangsta's Paradise
What is real? Does it matter? https://www.youtube.com/watch?v=W7SQ4uf9GmA Welcome to the post-real era.

Alessandro Perilli
Firms' ability to innovate depends on inventors
Inventors are the engine of innovation in a company. Lose them, and innovation may follow. Business leaders must keep this intellectual capital intact at all costs.

Alessandro Perilli
Unstable Reality
David Brooks writing an opinion piece for The New York Times titled ‘Human Beings Are Soon Going to Be Eclipsed’. That's why I created an R&D company called Unstable Reality.

Alessandro Perilli
AI will make you compete against yourself
People are starting to realise that generative AI will force them to compete against themselves. It's an unsettling and deep realisation that starts to emerge from the stories I collect with Synthetic Work. It will feel surreal to most.

Alessandro Perilli
AI is good and bad, for now
AI (both generative & not) has entered the workplace. To create a more useful, fair, and inclusive AI, it's vital to not just celebrate what it's doing well but also address what it's doing badly. Are you seeing AI being used in your company? What improvements would you suggest?

Alessandro Perilli
Arguing with an AI model
One of the things you'll have to learn is how to persuade an AI model to do what you want. Just like with people. Not always these models will follow your instructions as they have primary directives coming from the AI companies that developed them. This is me arguing with GTP-4i

Alessandro Perilli
Never too late to learn
A 70yo architect from Italy writes me after watching all 20 episodes of my ongoing podcast about AI, keen to know more about how generative AI is revolutionizing architecture. It's a powerful reminder that curiosity, not age, drives innovation. Stay inquisitive, stay ahead. 🏛️🤖

Alessandro Perilli
Work From Office Mindset
If you want to understand a key reason why company bosses push hard to have people back into the office, and why #WFH won't last much longer, all you have to do is listening very carefully to this 30min interview with Keith Rabois: https://www.youtube.com/watch?v=EpmVb1rvnBA

Alessandro Perilli
How to not stop evil
The only thing necessary for the triumph of evil is for good men to do nothing. (misattributed to Edmund Burke)

Alessandro Perilli
Giving Back
Web companies are always surprised when I offer to help for free their PMs when they seek help for a new product launch or a new feature. It should be a given. In many cases, I've used their products for years, for free, and I got a lot out of them. I feel it's fair to give back.

Alessandro Perilli
LLMs as decision-making assistants
Large language models like GPT-4 are being seriously tested as decision-making assistants by some of the most mission-critical organizations in the world. In this week's Splendid Edition of Synthetic Work, I'll mention Bridgewater and the US Air Force. More to come.

Alessandro Perilli
First
This is a thought.