Mathematica’s 35th anniversary! From a genius physicist to a "megalomaniac", the legendary experience of founder Wolfram

Shang En Mengchen from Ao Fei Temple
Qubit | Official account QbitAI

Hair and monitors are thinner, but code from 35 years ago still runs!

On the occasion of the 35th anniversary of the famous scientific computing software Mathematica , founder Stephen Wolfram released such a set of comparative photos with identical compositions, saying:

What was ahead of its time 35 years ago is still ahead of its time...

A large number of netizens lined up to express their blessings in the comment area, including many who use this software every day and have used it for 35 years .

In addition to Mathematica, which is usually used by professionals, Stephen Wolfram also designed the Wolfram Language and Wolfram Alpha, allowing people to visually complete high-end scientific calculations through natural language.

Just recently, he also wrote a popular science article explaining the principles of ChatGPT and compiled it into a book.

He built an empire of computing tools, and his products have helped countless researchers, engineers, and educators around the world.

Yet there was a time when he was criticized as much as he is praised now.

Some say he is a genius, others say he is a "megalomaniac."

And all of this started when he gave up theoretical physics and turned to studying cellular automata.

From genius physicist to egomaniac

At the age of 15 , Stephen Wolfram had already begun studying applied quantum field theory and particle physics.

At the age of 17, he entered Oxford University and completed his bachelor's degree and PhD in just 3 years. At the age of 20, he received a doctorate in theoretical physics from the California Institute of Technology.

During this period, he published many papers, one of which was on heavy quarks and was later widely cited in the academic community.

Later, while studying for a PhD at Caltech, Stephen Wolfram met the famous physicist Feynman .

Together they studied how to use cellular automata to simulate physical processes (such as turbulent fluid flow) and participated in the establishment of the discipline of complex systems.

He became the youngest recipient of the MacArthur Genius Grant with Feynman's recommendation letter, which was also one of the opportunities for him to later work at the Institute for Advanced Study (IAS) in Princeton.

But Wolfram's experience working in academia was unhappy, and he often didn't get along with his colleagues.

While at Princeton, his research interests slowly shifted from theoretical physics to complex systems, which he considered to be an emerging field outside the existing scientific system.

In 1985, he wrote to Feynman complaining about his situation and asking for guidance.

The treatment I received in IAS was getting worse and worse and I really had to leave. I want to build an institution to support my research into what interests me now.

Feynman's advice to him was to "do your own research."

You can't understand "ordinary people". To you, they are "dumb fools," so you won't tolerate them or treat their shortcomings with tolerance or patience.

At that time, Feynman believed that if Wolfram did management work, he would either drive himself crazy or be driven crazy by "ordinary people".

Minimize contact with non-technical people as much as possible, with one exception and that's falling madly in love! That's my advice, my friend.

After leaving IAS, Wolfram went to the University of Illinois at Urbana-Champaign, where he began developing the first version of Mathematica.

What started as a tool he designed for his own research later became determined to develop it into a general-purpose computing system that could be used by everyone.

In 1987, he left academia entirely and founded his own company, Wolfram Research, to continue the development and sales of Mathematica.

In the more than ten years since then, Wolfram's main label has gradually changed from "physicist" to "software tycoon" .

But at the same time, he has never given up research on complexity science.

Until 2002, he published a controversial book "A New Kind of Science", which attracted a lot of criticism in the scientific community.

In this book, he makes a very bold conclusion:

The universe is digital in nature and runs on what can be described as simple processes.

He also predicted that when the scientific community realizes this, it will have a major revolutionary impact on physics, chemistry, biology, and most other fields, which is why the book is titled like this.

However, physicist Freeman Dyson commented on this theory:

As scientists age, they often come up with grand, impossible theories. What's unusual about Wolfram is that he did it in his 40s.

This is a relatively polite comment among all, more like statistician Cosma Shalizi.

Shalizi believes that it will set the entire field back 10 years and that he will have to waste a lot of time correcting students who have read this book.

Wolfram has not been affected by these comments and has continued to develop his theory, and in 2020 claimed to have found a path to a basic physical theory...

Although not many people agree with his theory in the world of physics, mathematics or complex science.

But Mathematica, the tool he originally built to support his research, has become the most widely used software in science, engineering, and education.

From Mathematica to Wolfram Alpha

In 1986, just as personal computers began to have sufficient computing power.

At this time, if you want to complete different research on the computer, you need to use a variety of different software. Moreover, a new C language program must be created every time before starting a new physical operation, which is time-consuming and cumbersome.

Stephen Wolfram came up with the idea that maybe I could make a general computing system , and then I would just use this system, and other people would naturally think that this system is very practical.

On June 23, 1988, the first version of the computer algebra system "Mathematica" was released, causing a sensation.

Larry Smarr, director of the Institute of Communications and Information Technology at the University of California, San Diego, calls Mathematica software the most important scientific software ever created .

Before Mathematica, the field of complex technical computing could only be the domain of a few "master-level" computing experts.

All this changed rapidly with the emergence of Mathematica, which allows scientists, engineers and other professional groups to perform various complex mathematical operations as they wish, such as solving equations, finding derivatives, finding the inverse of matrices, drawing three-dimensional graphics, etc.

And more importantly, code that ran in 1988 can still run today.

When the first version of Mathematica was launched, Stephen Wolfram was still employed at the University of Illinois at Urbana-Champaign.

With the product selling well, the company is making good profits and is ready to go public.

But at that time, he decided not to go public because he wanted the company to always be his own, so that he could do what he wanted to do, such as long-term research projects.

Maybe you will ask why you can't make money even if you have money. In fact, he had an entrepreneurial experience before launching Mathematica, but the result was very unsatisfactory.

At that time, Stephen Wolfram, who was still a researcher at the California Institute of Technology, had access to the most advanced computers, but he continued to encounter failures.

The only way, he realized, was to build them himself.

So, I took action.

Because he knew little about business activities, he recruited an outside CEO to manage the company, while he continued to engage in his favorite scientific research. That is to say, from this moment on, things gradually developed in an uncontrollable direction.

According to his own later account, every time he wanted to offer advice, he was told that he didn't know much about business and should stay out of it.

It was precisely because of his previous bad experience in starting a company and because he always wanted to do something he liked that he decided to set up his own company, which is now Wolfram Research.

According to Stephen Wolfram, the world is representable, the world is computable, and the world is our knowledge. And now many search engines are just retrieving data, without trying to understand, analyze and calculate the data.

Therefore, he tried to make a high-end computing software. As long as you can describe what you want, the computer will do it for you.

In order to achieve this goal, he launched the computing knowledge search engine Wolfram Alpha, which was officially launched on May 18, 2009.

Wolfram Alpha's underlying operations and data processing are all implemented based on Mathematica, so Wolfram Alpha can answer a variety of mathematical questions and present the answers in a clear and beautiful graphical manner.

For example, when we enter 200g milk + 500g beef in the Google search box, what appears are food advertising images and content teaching you how to cook.

On WolframAlpha, the nutrition table of the corresponding food appears.

Compared with traditional search engines, WolframAlpha is more like an encyclopedia. For a variety of questions, it will not list related web pages with keywords, but "give direct answers."

Of course, these answers are not created out of thin air. The sources include major academic websites, Dow Jones, the U.S. Bureau of Investigation, etc.

At first glance, does it look like ChatGPT? Directly prompt you for the answer.

Although the common thread between the two is "natural language," and ChatGPT is excellent at automatically processing human-like events, not everything useful is so "human-like," he said.

Just like mathematical calculations, ChatGPT's performance has always been somewhat unsatisfactory. For a simple "chicken and rabbit in the same cage" question, the calculation results given by ChatGPT are often wrong.

In this regard, Stephen Wolfram believes that it will take a lot of energy to fundamentally solve this shortcoming of ChatGPT, but there is a shortcut, which is to make ChatGPT and WolframAlpha perfectly complementary.

By letting the latter's powerful structured computing capabilities and NLP understanding capabilities handle mathematical calculations.

Of course, this proposal was quickly adopted by OpenAI. Just this year, it was included in the first batch of plug-ins launched by ChatGPT.

In a recent interview, Stephen Wolfram revealed that he long ago hoped to democratize computing so that people would not need to rely on "priest-like agents" to obtain computing power.

In addition to developing plug-ins to connect his own software to ChatGPT, he is also trying to add large model capabilities to Wolfram Alpha.

At the same time, he also revealed other research directions being conducted, such as...

What exactly could the legend find in the cat in the party hat? Stay tuned ~

Reference links:
[1] https://twitter.com/stephen_wolfram/status/1672335093547819009
[2] https://mathcubic.org/article/article/index/id/85/cid/3.html
[3] https: //www.cantorsparadise.com/richard-feynmans-advice-to-a-young-stephen-wolfram-1985-d572dc360c18
[4] https://www.quora.com/What-do-mathematicians-think-about- Stephen-Wolframs-A-New-Kind-of-Science