- Tensor Black
- Posts
- The Video of Life
The Video of Life
Plus: Google and OpenAI strive for Her
Hello Tensor Black Fans!
The founding AI scientist at OpenAI quit, while OpenAI and Google flooded the press with a dizzying array of announcements on their new AI capabilities.
Abu Dhabi held their annual worldwide health conference, where every track mentioned AI in one form or another. We’re two weeks away from the annual meeting of ASCO in the US, where AI promises to be the hot topic of the day with over 5000 abstract submissions.
Level Up and get ready today:
New At-home Cancer Test
Predicting Cancer with Tensors
The Video of Life
Her… for Real?
Oncologist Shortage Crisis
Have a great week,
The TensorBlack Team
Google handed out a free, 45-page booklet on effective prompts at their Google I/O conference.
Make sure your Gemini prompts include:
Persona (who you are),
Task (what you want to do),
Context (background info), and
Format (how you want it to look)
Context includes relevant documents, videos and websites which are uploaded as attachments.
What’s Hot
A new at-home blood test device, Liberty, allows cancer patients to perform routine blood tests from home, reducing hospital visits and potentially increasing NHS productivity.
Details
The Liberty device received regulatory approval and is now being used at 12 NHS sites after successful trials at The Christie in Manchester.
Patients report significant reductions in stress and physical exhaustion from hospital visits, with simple, painless testing processes.
Though promising, the device's effectiveness and broader application require further research, with current trials involving limited participant numbers.
Why It Matters
Testing at home with inexpensive devices is a critical step towards reducing the cost of cancer prevention and oncology care.
We hope to see continued progress in this area, as we all witnessed during the pandemic for Covid19 tests!
Recently published in Nature, a novel framework integrates multi-omics data using autoencoders and tensor analysis to stratify cancer patients into precise risk groups, enhancing treatment strategies.
Details
Utilizes autoencoders for dimensionality reduction and captures non-linear relationships within ‘omics data.
Employs tensor analysis for feature integration, preserving significant multi-omics information across different cancer types.
Enables the classification of patients into distinct risk groups, significantly associating with patient survival rates and therapy effectiveness.
Why It Matters
This paper beautifully describes the AI technique behind phenomics (and video AI), perhaps the most advanced, data-driven approach to understanding cancer, detecting early transitions, and recommending actions.
Google DeepMind announced Veo, an advanced generative video model, transforms text prompts into high-quality, cinematic videos, enhancing creative storytelling and video production.
Details
Veo generates 1080p videos from text prompts, capturing cinematic effects like time lapses and aerial shots.
The model supports advanced editing capabilities such as masked editing and input-based video generation, aligning closely with user prompts.
Veo integrates technologies like latent diffusion transformers to maintain consistency and reduce flickering across video frames.
Why It Matters
Multimodal, multiomic, longitudinal patient records essentially form a patient’s “Video of Life.”
AI techniques that enable Hollywood movies are the same that enable early diagnosis, drug discovery, digital twins and treatment. Track their evolution closely.
Silicon Valley's latest AI technology promises to eliminate mundane tasks like writing emails or coding by introducing more interactive, possibly overly friendly virtual assistants — like Scarlett Johansson of “Her.”
Details
OpenAI and Google showcased new AI assistant developments, indicating a shift toward more intuitive and human-like interactions.
The new AI models, including OpenAI's GPT-4o and Google's Gemini, aim to transform everyday digital interactions and search functionalities.
Discussing potential impacts, tech leaders highlight the importance of addressing AI's capabilities responsibly, particularly its tendency to generate misleading information.
Why It Matters
AI is not an API or tool. AI is becoming a partner and collaborator in care, administration, and science.
Alberta faces a critical shortage of oncologists, endangering timely cancer care despite population growth.
Details
Alberta Medical Association reports the same number of oncologists as ten years ago, with 50 more needed soon.
Many trained doctors leave Alberta, worsening the shortage.
Delayed treatments result in long delays, poorer patient outcomes and increased palliative cases.
Why It Matters
Burnout is real: the increase in the cancer population, increasing complexity of care standards, and continual innovation are stretching oncologists to the limit. These demands will accelerate adoption of AI.