A rack of Tesla A100 from the Perlmutter supercomputer at NERSC/Berkeley Lab
Meanwhile, companies with large computing capabilities are making interesting forays in using AI for science, for instance Meta, which is developing OpenCatalyst in collaboration with Carnegie-Mellon University, where the goal is to create AI models to speed up the study of catalysts, which are generally very computer-intensive (see the Berkeley Lab Materials Project.) Now the cool part is to verify these results using x-ray diffraction at a synchrotron facilities. Something a little similar happened with AlphaFold where newly derived structure may need to be tested with x-rays at the Advanced Light Source: Deep-Learning AI Program Accurately Predicts Key Rotavirus Protein Fold (ALS News)
Last week I was lucky to meet with Vanessa Chan, the Chief Commercialization Officer for the Department of Energy and Director of the Office of Technology Transitions. She wanted to hear what kind of hurdles when it comes to start a company (hint: a lot.) I told her that a major, overlooked issue is that you generally to be a permanent resident to start at company in the US, whereas two-thirds of postdocs are foreign nationals and on visas. There are ways to get around the requirement (such as Unshackled), but it’s a little sad not more is done to provide support to those willing and able (plus – it is a well-known trope that many US companies are founded by foreign nationals, what I tend to believe is among what sets California apart from other states and other countries, where entrepreneurship doesn’t flourish as much as expected despite many efforts)
I am a native French speaker, and I have always been confused by the ubiquity of English, language which is actually quite difficult to speak (why is tough, though, thought and enough so different?) And I was also puzzled the difference between liberty and freedom – no one could ever explain me the difference, even though “Freedom” is probable the most overused concept in American society (French has “Liberté” in its national motto, but is has nothing to do with “free” as in “free sample.”)
Finally, I found an interesting explanation by Jorge Luis Borges, who sees this as a feature, not a bug:
I have done most of my reading in English. I find English a far finer language than Spanish.
Firstly, English is both a Germanic and a Latin language. Those two registers—for any idea you take, you have two words. Those words will not mean exactly the same. For example if I say “regal” that is not exactly the same thing as saying “kingly.” Or if I say “fraternal” that is not the same as saying “brotherly.” Or “dark” and “obscure.” Those words are different. It would make all the difference—speaking for example—the Holy Spirit, it would make all the difference in the world in a poem if I wrote about the Holy Spirit or I wrote the Holy Ghost, since “ghost” is a fine, dark Saxon word, but “spirit” is a light Latin word. Then there is another reason.
The reason is that I think that, of all languages, English is the most physical of all languages. You can, for example, say “He loomed over.” You can’t very well say that in Spanish.
And then you have, in English, you can do almost anything with verbs and prepositions. For example, to “laugh off,” to “dream away.” Those things can’t be said in Spanish. To “live down” something, to “live up to” something—you can’t say those things in Spanish. They can’t be said. Or really in any Roman language.
I really enjoy this notion of physicality – onomatopoeia are a vibrant part of the language: whisper, gulp, slam, rumble, slushy, etc.
Wovon man nicht sprechen kann, darüber muss man schweigen.
Whereof one cannot speak, thereof one must be silent. – Ludwig WittgensteinLet’s all clap for Borges!
Lately, there’s been quite some interest in LLM, thanks to ChatGPT
Write two paragraphs in the style of Tim Ferriss in the style of The Four-hour Workweek explaining on how to generate passive income with ChatGPT
Are you looking for a way to generate passive income with ChatGPT? Then look no further than affiliate marketing. By promoting ChatGPT’s products or services to your audience and earning a commission on any sales you refer, you can create a reliable and scalable source of passive income. Whether you have a blog, a podcast, or a social media following, affiliate marketing with ChatGPT is a low-risk, high-reward way to monetize your online presence.
But that’s not all. Another way to generate passive income with ChatGPT is by creating and selling digital products, such as e-books or online courses, related to the topics covered by ChatGPT. By leveraging your expertise and knowledge, you can create a valuable resource for your audience while also generating passive income. And with platforms like Udemy and Teachable, it’s never been easier to create and sell digital products online. So why not take advantage of ChatGPT’s wealth of knowledge and expertise and start generating passive income today?
Plagiarism
One of the big concern du jour is plagiarism. Here’s an interesting piece by Noam Chomsky in the
The human mind is a surprisingly efficient and even elegant system that operates with small amounts of information; it seeks not to infer brute correlations among data points but to create explanations.
And a more elaborate conversation:
Emergence
However, it seems there is a little more at stake. A former classmate of mine who works at Microsoft talked about sparks of Artificial General Intelligence in a podcast – there are emergent properties in AI, meaning that it does not only remix text, but also seem to have an ability to create things merely from structure (short of being able to think.)
ChatGPT 4.0 can generate code to make a unicorn, even though it has no notion of what a unicorn is, and that code is procedural. (excerpt from Bubeck’s lecture atBerkeley on sparks of AGI)
Physics of AI
The same friend is actually exploring what he dubs the Physics of AI: trying not to understand why AI works, but under what circumstances you can make it work.
The key insights (nevermind the jab at physicists!) is that there are weird phase transitions at work: depending on how you pick your learning parameters, you can get wildly different results – turning the teaching of the machines into an art itself.(update March 2024: check out this fascinating analysis from Sohl Dickstein on how Neural network training makes beautiful fractals)
In the end, humans will be only be there for liability. In a way this is how American society has evolved.
ChatGPT as anonymous expert
Personally, I think that the ability to get answers to question without the fear judgement – asking dumb or intimate questions – will have an incredible impact, perhaps more than the productivity gains that are touted by experts (sure – you can generate reports with AI, then summarize these reports with AI: perhaps the thought should have been boiled down i. the first place.)
“To ask the right question is harder than to answer it.”
–Georg Cantor
Yesterday I invited Tanya Zimbardo from the San Francisco Museum of Modern Art to give a talk at Berkeley Lab (details about the even can be found here: Hybrid Forms: Connecting Art and Science)
Tanya Zimbardo (SFMONA) at Berkeley Lab
It was quite interesting to hear her perspective on a topic which is close to my heart, and happy to hear many references to Rafael Lozano-Hemmer, who currently has the Techs-Mechs exhibition running at the Gray Area, but also quite surprising not hear anything about Jim Campbell (whose art glows atop the Salesforce building “Eye of Sauron”) or the work of Illuminate.
In a our last paper, we present the use of machine learning to get the most of x-ray adaptive optics – and it works like magic! This was a great work accomplished by Gautam Gunjala, a grad student from UC Berkeley under a SCGSR grant, together with our wonderful colleagues from the Advanced Photon Source.
X-ray adaptive mirrors are very nice, because they allow to correct the shape of x-ray beams, when the beam gets distorted by mirror deformation or misalignment. That’s why we want to use them in the latest generation of synchrotron light source such as ALS-U or APS-U.
Maria Zurek (left) and I (right) with Jeff Welser (IBM VP of research, center) and the rest of the Berkeley Lab Postdoc Association in October 2019
She gave a few great pieces of advice, here are my favorites:
– come to the conference with your slides READY. If they’re not ready, you will miss most of the social hours, where the networking happens (that’s a hard one – but very true: a conference is mostly about meeting people, not presenting your research.)
I am very happy to announce that was granted en Early Career Research Award by the US Department of Energy, to work on the DREAM beam: Diffraction-limited Radiation Enhancement with Adaptive Mirrors for x-ray coherent beamlines.
There isn’t much representation of scientists in popular culture, with a only few movies standing out, such as a “A Beautiful Mind” (on John Nash) or “Good Will Hunting.” There’s been a few more in the biopic genre lately, such as the “Imitation Game” on Alan Turing or “The Theory of Everything” on Stephen Hawking, and soon a movie on Robert Oppenheimer by Chris Nolan.
But the representation of women in science and technology is even less frequent. Things seem to be changing, and during the pandemic there’s been a few biopics on women scientists, to which I want to bring attention to:
Secrets of the Surface, a movie on Maryam Mirzakhani, the only women ever to receive the Fields medal
Radioactive, a movie about Marie Curie, with Rosamund Pike
All of them have been deprived of a theatrical release, and I find it a bit sad they haven’t been delayed, but perhaps there’s been increased distribution through streaming platforms.
I should also mention slightly older movies such as “Hidden Figures“, “Contact“, “Arrival” and “Interstellar” – surprisingly all about space exploration.
Why can’t we see beakers, petri dishes and lasers?
