Dawn of a New Era

Plus: AI's Power Problem

Hello Tensor Black Fans!

Microsoft released some new AI PCs, where continuous screenshots are a bit invasive, but interesting features nevertheless. OpenAI may have been trying too much to be like Her. With Google’s 2 trillion searches a year, achieving an 0.00001% error rate is still a hiccup every 3 minutes.

We’re in the early days, where GenAI experiments will have growing pains.

What stood out last week? Max Tegmark. His talk about AI as the dawn of a new species inspired us! Longterm, AI is not an API.

Level Up and more:

  • AI: Dawn of a New Era

  • Loneliness Raises Cancer Risk

  • Cosmic Microbiology

  • AI’s Power Problem

  • Building Trust in AI

Have a great week,
The TensorBlack Team

Writing a post each week on LinkedIn is a great way to raise your profile in the market, while also helping your patients in their daily lives. Try the following for inspiration! See an example here.

Create a series of educational blog posts for LinkedIn targeting individuals living with [condition]. Develop 3 content topics, each with a brief description limited to 200 words. Each title needs to be short, pithy, informative, 3-4 words and hook the reader to learn more. Use a professional tone, informative, and follow the writing of newspapers where the title is catchy, the first sentence gives you the gist of the article and serves as a hook to read more, with 2-3 more sentences giving the key takeaways. Provide a recent article (title, source) I can cite and find with Google.

— TensorBlack Team

What’s Hot

Max Tegmark of MIT explores the unexpected advancements in AI technology, comparing it to the dawn of a new species, and its profound implications for various fields, including oncology.

Details:

  • AI’s Accelerated Evolution: Despite previous expectations, AI models like ChatGPT-4 have advanced rapidly, achieving language mastery without fully understanding the human brain, paralleling early breakthroughs in flight.

  • Beyond Transformers: Current AI architectures, such as transformers, are likened to the vacuum tubes of computing, suggesting future models will be more efficient, requiring less data and power.

  • AI as Collaborative Agents: AI is transitioning from being a mere tool to becoming autonomous agents capable of complex reasoning and actions, similar to human cognition.

Why it Matters:

AI systems will soon autonomously analyze vast datasets, generating insights quickly. Oncologists can expect more precise diagnostics and personalized treatment plans. Understanding AI will be crucial for staying at the forefront of medical innovation.

Higher levels of loneliness significantly increase the mortality risk among cancer survivors, highlighting the need for targeted support programs.

Details

  • The study found that cancer survivors with severe loneliness had a 67% higher mortality risk compared to those with low or no loneliness.

  • Data from the Health and Retirement Study (2008-2018) included 3,447 cancer survivors aged 50 and above, with follow-up through 2020.

  • Loneliness was measured using an 11-item UCLA Loneliness Scale, and results showed a dose-response relationship between loneliness levels and mortality risks.

Why It Matters

Loneliness sucks. Screen for loneliness in routine oncology and primary care, which can lead to timely interventions and improved overall health for survivors.

Visual similarities between microscopic organisms and cosmic structures reveal the underlying physics and chemistry of our universe.

Details

  • Microscopes and telescopes show how basic physics and chemistry are present in both tiny organisms and massive galactic structures.

  • Images of filiform papillae on a rabbit's tongue resemble Jupiter's swirling cyclones, and Raji cells parallel the sun's surface.

  • Coronavirus spike proteins are visually similar to cosmic debris in a supernova, while colors in a mouse's eye resemble Saturn's vortex.

Why It Matters

Life at a nanoscale is reflected at an astronomical level. That’s just cool.

Generative AI's growing energy demands are straining electricity grids, prompting concerns over sustainability and efficiency.

Details

  • Generative AI systems use significantly more energy compared to traditional software, with estimates suggesting 33 times more energy consumption.

  • Data centers, which power generative AI, are expected to double their electricity usage by 2026, equating to the consumption of a country like Japan.

  • Innovations in AI hardware, such as Nvidia's Grace Blackwell chips, aim to improve efficiency but still contribute to substantial energy demands.

Why It Matters

Once we get through the NVidia GPU squeeze, raw power is next. Expect to see Energy Star ratings on AI models in the future and a “land grab” for power contracts.

Establishing clear guidelines and fostering open discussions can make AI more trustworthy and effective in organizations.

Details

  • AI efforts at Aramark include a task force with regulatory and exploratory committees to manage risk and gather ideas.

  • Key guidelines: avoid inputting sensitive data and always use human judgment alongside AI outputs.

  • The company encourages employee involvement through AI ambassadors and regular idea submissions.

Why It Matters

Understanding and establishing trust in AI is crucial for oncologists and caregivers who are integrating AI tools into their workflows.

Clear policies and open communication can help mitigate risks and enhance the effectiveness of AI in healthcare.