- The Startup Breakdown
- Posts
- Glass: AI Revolutionizing Clinical Decision Making
Glass: AI Revolutionizing Clinical Decision Making
Google for Startups Accelerator: AI First Demo Day
This is The Startup Breakdown, the only newsletter looking into the new Y Combinator companies that you need to know. By joining this growing community of hundreds of founders, investors, and other startup aficionados (think i spelled that right?), you're getting a firsthand look at the companies that may be the next Airbnb, Coinbase, or Reddit.
If you'd like to receive these newsletters directly in your inbox once a week, hit subscribe and never miss an email!
Premium subscribers include investors and YC-backed founders looking to stay ahead of the latest startup news.
AI in Healthcare - Fundamental Change to the Diagnostic and Treatment Process
Addressing Administrative Overload in Healthcare
THE PROBLEM
How many hours a day do you spend checking emails, scheduling appointments, or answering texts?
Chances are, quite a few.
Our med school friends experience this, too, though unlike us normal folks missing out on an extra episode of Dated and Related, every minute that a doctor spends on administrative tasks is a minute they can’t spend treating patients.
On average, the time a medical professional spends on these mundane tasks amounts to an entire workday.
One of the biggest problems that I saw was that doctors truly didn’t have good software to support their critical clinical workflows and making complex decisions for very sick patients.
Introducing Glass AI for Healthcare
THE SOLUTION
Glass, an AI platform transforming healthcare, is addressing this, empowering clinicians with differential diagnoses and clinical planning.
Doctors can type (thank goodness they don’t write amirite?) the patient’s details (age, gender, etc) and symptoms and get instant diagnoses to test for.
Then, once the doctor gives the patient an actual lookover and confirms the diagnosis, the AI automatically generates a plan of attack for treatment.
Glass AI's New Odyssey Product
Even better, as of yesterday, they have officially launched their new Odyssey product for EHR integration (talk about early to know 👀).
The tool actually listens in on the conversation, taking notes and combining with the patient’s existing medical record (and subsequently the diagnosis) in real time, only further boosting clinician efficiency.
THE NUMBERS
Glass AI by the Numbers
Traction:
70K users
480K queries
Raised from YC & Breyer Capital (pre-seed), Initialized ($5 million seed)
Market Size:
TAM: $16.85 billion
SAM: $5.055 billion
SOM: $252.75 million
Competition:
DeepScribe, Notable
Ambience AutoScribe, Babylon Health
IBM Watson Health
The Team Behind Glass AI
Team:
Dereck Paul, MD, CEO: internal medicine resident at Brigham and Women's Hospital, Harvard Medical School and a medical student at the UCSF School of Medicine
Graham Ramsey, CTO: designed the first telemedicine-specific EHR at PlushCare, and helped build the world's best fertility dashboard & ovulation tracking app at Modern Fertility
Risks and Challenges for Glass AI
Risks:
Data security: Glass deals with a ton of highly sensitive patient data, and the potential consequences of a breach (and the regulatory scrutiny the company is likely to experience as a result of this potential) are devastating to the company and to the space
Differentiation: My first thought when looking at the product was “how is this different from ChatGPT?” The traction eased my concerns a bit, but the product still has a long way to go prove to be anything more than a “ChatGPT wrapper”
Admittedly, this opinion is a bit skewed by my familiarity with the AI space. it’s possible that for medical professionals without such close exposure, everything about Glass Health is completely novel
Adoption risk: also eased by the company’s traction numbers, but healthcare is a notoriously hard
It is notoriously difficult.
We do something very unusual in healthcare which is that we are a Product-Led Growth company… Those 70,000 doctors who signed up did so organically without any marketing spend, mostly from word of mouth.
What I Like About Glass AI
What I like:
Advisors: part of the reason for this early traction is the strong network of more than a dozen highly connected medical professionals on the advisory board, leveraging their networks to get the software into the hands of as many doctors as possible
Augment vs. replace: a faulty strategy wielded by other companies attempting to solve this and similar problems is attempting to automate out the professional… not only is this error-prone, but it’s nearly impossible to do given the pushback and lack of feedback the founding team will get from the niche group of people their product is attempting to infiltrate
Simplicity: this is pretty much a ChatGPT wrapper, with a single box for entry and a clean, simple response and follow up design. Glass knows their audience, and they’re building the product that they want to use, not what the tech world (see: me) thinks they should build
Opportunities for Glass AI's Growth
Opportunities:
Build own LLM: Glass is about to have access to a massive amount of medical data, incessantly reinforced and corrected by tens of thousands of medical professionals on a daily basis… the finetuning that the company will be able to perform could make for the strongest medical diagnostic and treatment models in the game
Data licensing: What to do with all of this data and the model? Why, open up a new revenue stream of course! Glass can consider selling encrypted and redacted data to medical researchers for the sake of treatment development, and even better, by leveraging this position of power, it can organize partnerships with treatments to promote their products to the Glass network of doctors
Verticalization: Glass’ new Odyssey product is a massive step in the direction of verticalizing their product offerings for the entire scope of the medical examination and treatment value chain, but there’s still a ton of room to go, from the pre-examination data collection and scheduling process to the prescription fulfillment and monitoring stages
LINKS OF THE WEEK
Share this newsletter with one friend to get access to the Links of the Week section in every newsletter, including today’s! 👇️
BeReal was acquired by Voodoo, a French app developer, for ~$530 million
This was probably the best outcome for the founding team given the lack of a clear path to monetization… just look at how much Snap has struggled and the stagnant user growth numbers
It is even reported that the CEO had been exploring a sale last year, so he was clearly aware that the peak had already been reached… frankly, I’m amazed that they were able to bring in this much
A theme in today's Links section: wins for European startups…
BeReal was acquired by Voodoo (French mobile app & gaming company) for €500M
I want to know who BeReal's bankers are... seems like they did a great job of maximizing value.
Download numbers over the last 2 years:
— Sheel Mohnot (@pitdesi)
3:49 PM • Jun 11, 2024
Rasberry Pi, the people’s computer, went public on Tuesday, shooting up 32% in its initial day of trading. Takeaways…
Good get for the London Stock Exchange which sees the vast majority of the UK’s tech companies list on the NYSE, opting for buttermilk biscuits over those powdery ones the Brits eat with their tea
The success is a bit shocking considering the use cases for the company’s cheap, small computers are mostly limited to niche hobbyist ventures
Loyal users aren’t thrilled as they see this as a pivot away from community-powered products towards catering to shareholder demands, meaning cost cutting and low quality in pursuit of ~profits~
Relatedly, ARM, a major strategic investor in the company, has stated its interest in increasing its stake
Twitter will finally be hiding all user likes
Elon has long voiced his opposition to users’ likes being public and visible on their profiles, and after an initial rollout with Premium subscribers, the privacy enhancement officially began public rollout earlier this week
He claims that he is tired of people being attacked for the posts they like, so he wants people to be able to give their thumbs up in peace
I actually agree with this, and I think it’s stupid that something as simple as a little button indicating “good tweet” can get someone virtually persecuted, especially as someone who likes just as many tweets as a way of saying “this is the dumbest thing on the internet right now,” but I must admit - I will miss the days of Ted Cruz getting caught like porn
Joey Chestnut was banned from this year’s Hot Dog Eating Contest after announcing a partnership with Impossible Foods
A spokesperson for Nathan’s claims that the greatest athlete on the face of this Earth had broken an exclusivity agreement that had linked the man, the myth, the legend and the sponsor of the contest, meaning he is unlikely to defend his title and win his 18th very elastic, very loose belt
Jokes aside, this genuinely might have a negative impact on the event which has become a legitimate revenue stream
The power of the creator economy is only beginning, and no industry is safe
What the actual fuck
DeepMind and Harvard just created a virtual rat powered by an AI neural network.
It's capable of mimicking agile movements and neural activity of real-life rats — massively opening up new research with testing on AI animals, and expanding robotics
— Rowan Cheung (@rowancheung)
6:48 AM • Jun 14, 2024
|
Last word 👋
How am I doing?
I love hearing from readers, and I’m always looking for feedback. How am I doing with The Startup Breakdown? Is there anything you’d like to see more or less of? Which aspects of the newsletter do you like most?
How disappointed would you be if you could no longer read this newsletter? |
Cheers to another day,
Reply