By MIKE MAGEE
Let me be the primary to introduce you to Claude Elwood Shannon. In case you have by no means heard of him however think about your self knowledgeable and engaged, together with on the interface of AI and Drugs, don’t be embarrassed. I taught a semester of “AI and Medicine” in 2024 and solely not too long ago was launched to “Claude.”
Let’s start with the truth that the product, Claude, shouldn’t be the identical because the particular person, Claude. The particular person died 1 / 4 century in the past and aside from these deep within the area of AI has largely been forgotten – till now.
Amongst these within the know, Claude Elwood Shannon is sometimes called the “father of data concept.” He graduated from the College of Michigan in 1936 the place he majored in electrical engineering and arithmetic. At 21, as a Grasp’s scholar at MIT, he wrote a Grasp’s Thesis titled “A Symbolic Evaluation Relay and Switching Circuits” which these within the know declare was “the birth certificate of the digital revolution,” incomes him the Alfred Noble Prize in 1939 (No, not that Nobel Prize).
None of this was notably apparent in these early years. A University of Michigan biopic claims, “Should you had been searching for world changers within the U-M class of 1936, you most likely wouldn’t have singled out Claude Shannon. The shy, stick-thin younger man from Gaylord, Michigan, had a studious air and, at occasions, a playful smirk—however not one of the apparent facets of greatness. Within the Michiganensian yearbook, Shannon is yet one more face within the crowd, his tie tightly knotted and his hair neatly parted for his senior picture.”
However that was one of many historic misreads of all time, according to his alma mater. “That unassuming senior would go on to take his place among the many most influential Michigan alumni of all time—and among the many towering scientific geniuses of the twentieth century…It was Shannon who created the “bit,” the primary goal measurement of the data content material of any message—however that assertion minimizes his contributions. It will be extra correct to say that Claude Shannon invented the trendy idea of data. Scientific American referred to as his groundbreaking 1948 paper, “A Mathematical Theory of Communication,” the “Magna Carta of the Information Age.”
I used to be launched to “Claude” simply 5 days in the past by Washington Put up Expertise Columnist, Geoffrey Fowler – Claude the product, not the particular person. His article, titled “5 AI bots took our tough reading test. One was smartest — and it wasn’t ChatGPT,” caught my eye. As he defined, “We challenged AI helpers to decode authorized contracts, simplify medical analysis, speed-read a novel and make sense of Trump speeches.”
Judging the outcomes of the medical analysis check was Scripps Analysis Translational Institute luminary, Eric Topol. The 5 AI merchandise had been requested 115 questions on the content material of two scientific analysis papers : Three-year outcomes of post-acute sequelae of COVID-19 and Retinal Optical Coherence Tomography Features Associated With Incident and Prevalent Parkinson Disease.
To not bury the lead, Claude – the product – received decisively, not solely in science but additionally general in opposition to 4 identify model rivals I used to be acquainted with – Google’s Gemini, Open AI’s ChatGPT, Microsoft Copilot, and MetaAI. Which left me a bit embarrassed. How had I by no means heard of Claude the product?
For the reply, let’s retrace a little bit of AI historical past.
The New York Times headline in 2023 introduced the speedy progress of generative AI as “Thrilling and Scary” after 4 years of monitoring its’ progress. Their expertise columnist wrote, “What we see rising are machines that know purpose, are adept in any respect human languages, and are in a position to understand and work together with the bodily setting.”
Leonid Zhukov, Ph.D, director of the Boston Consulting Group’s (BCG) International AI Institute, believed then that choices like ChatGPT-4 and Genesis (Google’s AI competitor) “have the potential to turn into the brains of autonomous brokers—which don’t simply sense but additionally act on their setting—within the subsequent 3 to five years. This might pave the best way for absolutely automated workflows.”
OpenAI co-founders Elon Musk and Sam Altman in 2016 initially expressed considerations about machines that not solely mastered language, however might additionally suppose and really feel in super-human methods. Wishes for security and regulatory oversight linked them in these early years. However that didn’t final for lengthy. When Musk’s makes an attempt to achieve majority management of the now profitable OpenAI failed, he jumped ship and later launched his personal enterprise referred to as “XAI.”
Within the meantime, the Open AI Board staged a coup, throwing Sam Altman over-board claiming he was now not into regulation however slightly all in on an AI profit-seeking “arms race.” That solely lasted a number of days, earlier than Microsoft, with $10 billion in hand, positioned Sam again on the throne. Within the meantime, Google engineers, who had been credited with the unique break- by algorithms in 2016, created Genesis, and the complete blown arms race was on, now together with Fb with it’s MetaAI with super-powered goggles.
Altman later penned an op-ed titled “The Intelligence Age” wherein he defined, “Expertise introduced us from the Stone Age to the Agricultural Age after which to the Industrial Age. From right here, the trail to the (AI enabled) Intelligence Age is paved with compute, power, and human will.”
Claude was born that very same yr. Its’ dad and mom had been sibling co-founders of the 2021 public-benefit company, Anthropic, Dario Amodeo and Daniela Amodeo. They had been the VP of Analysis and the VP of Security & Coverage at OpenAI till Sam Altman’s conversion of that non-profit right into a capped for-profit entity (with Microsoft within the wings) created excessive ranges of rigidity and mistrust in higher ranks who felt security and public-good had been compromised. The entire concept, in spite of everything, was for OpenAI to “construct safeAI and share the advantages with the world.”
In December 2020, Dario, Daniela and 14 different OpenAI researchers jumped ship. Their new Board endorsed a twin mission to: search revenue for shareholders as a part of their fiduciary duty,” whereas creating “transformative AI that helps individuals and society flourish” and if want be “pursue AI security and ethics over creating revenue.” Their method to “useful and innocent” AI assistants was anchored in a dedication to “Constitutional AI” on their 1 yr anniversary in 2022. This human creation (the AI Structure) juries the boundaries of usefulness and security. Within the picture of Claude Elwood Shannon, Claude, the AI with a soul, was born.
They embraced a way for growth referred to as “Reinforcement Learning from Human Feedback” (RLHF). Definition: RLHF = “Fashions engage in open-ended conversations with human assistants, producing a number of responses for every enter immediate. The human then chooses the response they discovered most useful and/or innocent, rewarding the mannequin for both trait over time.” This allowed Anthropic to “engage fashions in open-ended conversations with human assistants, producing a number of responses for every enter immediate.
As the method advanced, they had been in a position to prepare the AI to grade the AI on consistency to the Structure that they had established. The AI now was in a position to grade itself harmlessness and helpfulness. The brand new course of, Reinforcement Learning from AI Feedback (RLAIF) was now the automated decide of RLHF. Going one step additional, Dario disclosed the Structure which bolstered Anthropic’s dedication to transparency and public service.
What can Claude do? In can generate textual content in tandem, summarize, search, code, and extra with excessive accuracy because it doesn’t depend on Web seek for content material. Researchers are actually quick at work coaching Claude to “generate responses primarily based on character traits …like curiosity, open-mindedness, and thoughtfulness.”
Apart from profitable the Washington Put up liberal arts check (together with legislation, medication, literature and politics), the Claude web site (with free entry) had 100 million visits in March 2025, and its iOS app had 150,000 downloads inside its first week of launch in Might, 2024. Anthropic has raised $18.2 billion as of Might, 2025, (2nd solely to OpenAI) with Amazon as its prime investor at $8 billion in return for naming its cloud service (AWS) its major cloud and coaching accomplice. Google is in as properly at $2 billion.
On June 5, 2025, Dario penned an opEd within the New York Occasions titled “Don’t Let A.I. Companies off the Hook.” In it he argues aggressively for a concentrate on transparency stating “That is about responding in a clever and balanced option to extraordinary occasions.” One can nearly see Claude Elwood Shannon within the shadows, quietly smiling.
Mike Magee MD is a Medical Historian and an everyday contributor to THCB. He’s the creator of CODE BLUE: Inside America’s Medical Industrial Complex. (Grove/2020)
Source link
