“We are in the early days of AI”

Elad Gil has seen it all. This week he poured his experience as a successful startup founder, employee at Google and Twitter, and now venture capitalist, to predict the shape of things to come in the AI sector. Listen:

These are still early days. The Cambrian Explosion of AI started mere months ago. Everything prior to that was prologue.

There will be 4 waves of AI adoption. (We’re already in Wave 2). The first wave was for companies already in the generative AI space (“AI natives”) such as OpenAI, Midjourney, Stable Diffusion, etc. The second (current) wave is made up of early-adopter startups (Perplexity, Langchain) and fast-moving incumbents (Notion, Adobe, Microsoft). The third wave (coming soon) will be powered by a huge crop of innovative new startups, and will be quite exciting for all. The fourth wave (2024? 2025?) will see the first big enterprise adopters who are slower to build products, but have potentially massive reach.

The power and quality of AI will only continue to increase. AI systems are already equal to or better than human experts in many applications. Google’s MedPalm 2 model gives better answers to medical questions than physicians do, as rated by physicians. The model is so good, that input from physician experts make it perform worse (?!?)

Read the original blog post by Elad Gil.

IBM study predicts 40% of workers will need to be reskilled by 2027

A new IBM study polled 3,000 C-suite business leaders and 21,000 workers globally, and found that there was a consensus that AI would make a massive impact in the near term on the content of workers’ jobs, requiring retraining for up to 40% of the global workforce. The chart below illustrates that the most valued skill in 2016 (STEM Proficiency) is now the least valued skill in 2023. Similarly, the most valued skill in 2023 (Time Management and the Ability to Prioritize) was only a middling-rated skill in 2016.

IBM study says 40% of workers will need to be reskilled for AI

Judge rules that AI cannot own copyright

This ruling has been widely mis-reported. The judge did NOT rule that works created with AI cannot be copyrighted. He ruled that an AI cannot own the copyright. Only people who create works - however artificially assisted - can claim copyright. The analogy was made to photography - the camera takes the picture, but the person makes the esthetic decisions. So by that logic, if a person crafts a killer prompt, and chooses which version of the output is to be kept, that person should be able to claim copyright for that output. We’ll see if the courts agree in future cases.

This ruling was the product of a bizarre case in which an artist tried to claim that the AI he used was the actual author of the work, and should be granted copyright in its own name.

Judge rules against AI on copyright

Neuroscientists recreate a Pink Floyd song from brain electrodes

Scientists at UC Berkely decoded the brain activity of neurosurgical patients by putting electrodes directly on the surface of the patients’ brains, and recording the electrical activity while each patient was listening to Pink Floyd’s “Another Brick in the Wall.” They were then able to recreate the musical clip from the AI-decoded brain waves. Click on the link below and judge for yourself how faithful the recreated sound clip is to the original.

Neuroscientists recreate Pink Floyd song from brain activity

Here’s the original paper

Snapchat’s MyAI goes rogue, posts bizarre clip - here

YouTube partners with Universal Music Group on AI Music Incubator - here

Microsoft embeds Python into Excel - here

Google releases paper on how to get LLMs to train themselves - here

Chinese launch satellite with AI-powered brain - here

AI bots beat humans (by a lot) at solving bot-spotting CAPTCHAs - here

Nvidia is booming once again.

After Nvidia’s earnings report was released this week, their stock increased by 9% after-hours this Wednesday. In the report, their year-over-year sales growth was 101%, and this quarter’s revenue was $13.5 billion dollars (a new record for the company). Just this year Nvidia’s stocks have climbed 220% and the demand for their product keeps increasing. Now they are planning to do a $25 billion dollar buy back of their shares because currently the company thinks their stock is undervalued. Is this the right move?

Zoom is jumping into the AI game, and promises of new features are on the horizon.

Zoom’s stock also jumped after-hours after the company said they are expecting higher earnings this year. Now they are promising the use of AI in their platform expanding on the ‘heartfelt’ notion of trust as Yuan, the CEO of Zoom says, “we strongly believe that technology should advance trust” (I’m crying). Is this believable considering the stunt they tried to pull last month on their terms and conditions? Who knows.

CNN Business: Zoom shares jump as CEO says company to develop and deploy AI tools.

The AI boom is hurting other stocks. Big companies are worried about their shares.

AI has officially transferred into the stock pool and they are no longer little fish. The pre-existing big fish are worried. Jim Cramer, a co-anchor at CNBC, spoke about the “recession stocks” that are slowly losing popularity. For instance, Campbell Soup has dropped 19% over the past three months and the S&P 500 has declined by 3%. People are buying into the “Magnificent Seven” referring to the top tech stocks this year such as: Nvidia, Meta Platforms, Microsoft, etc. Will this shift effect companies like Campbell more drastically over the next year?

CNBC: Cramer says AI hype is hurting this group of stocks that should benefit from recession worries.

China isn’t looking too good. What will happen to the adjacent economies?

China, one of the big economies of this era, is dealing with a major real estate crisis that might send shockwaves through the economy. Could this be a repeat of the Lehman Brothers in 2008 (one of the US investment banks that crashed in the recession)? Xiaoxi Zhang, a financial analyst at Gavekal Research, says “The worry is that the ‘Lehman moment’ beckons, threatening the solvency of China’s financial system.” She then goes onto saying that Beijing’s “regulatory vigilance” makes an outcome unlikely. Despite the trust in China’s financial infrastructure, Wall Street observers are concerned that the crash might cause a domino effect across the globe.

Le Monde: China's real estate crisis creates limited risk for the global economy

University of Michigan takes a huge leap into AI

U-M has downloaded AI onto their servers, emphasizing privacy, accessibility and equality for the university. With ‘initially’ a no-cost system it has generous usage limits for students, faculty and staff. U-M’s GPT works seamlessly with screen-readers, which do not work with other AI tools like ChatGPT, and all the data is is private and will not be used to train its AI models. To build this platform the school’s ITS collaborated with Microsoft and has launched three AI services that will be accessible across Ann Arbor, Flint, Dearborn and Michigan Medicine. U-M’s president stated “Using these tools responsibly, I am confident that our community of scholars will make a positive difference in the world.” It’s great, until you have to pay.

Michigan News: U-M debuts generative AI services for campus

Hot Take From a College Student: Rise of Social polarity that might be caused by AI.

By Addie Politi

As AI increasingly integrates into the lives of college students, it has the potential to amplify existing social disparities. Despite the challenges posed by COVID-19, the college social scene was on an upward trend. Students were pushing past increased social anxieties because of COVID isolation. However, this might all change due to the AI technologies that have aided students’ education, and will continue to improve.

As AI becomes a more valuable tool for students, there is a possibility that traditional study groups, which have historically played a crucial role in forging lifelong friendships, could become a thing of the past. The ample gatherings for homework assignments or group projects no longer will be an excuse to hang out with friends. AI might become the “friend that knows the answers” or the “friend that is good at explaining things.”

As a college student, I have noticed that there has already been a set back from COVID-19 and the social scene can be less lively, but I am worried that AI will completely take away the foundations of socialization in education.

AI vs. Students: ChatGPT, Notion AI and Others’ Impact on Learning Experience

Examining the Impact of Artificial Intelligence and Social and Computer Anxiety in E-Learning Settings: Students’ Perceptions at the University Level