Detecting Cancer From a Drop of Blood (The Anti-Theranos) 

Here’s some good further reading from Johns Hopkins (specifically for lung cancer 99.8% accurate): www.hopkinsmedicine.org/news/newsroom/news-releases/2024/06/artificial-intelligence-blood-test-provides-a-reliable-way-to-identify-lung-cancer

She wasn't "Ahead of her time" she was a fraud.

Elizabeth Holmes from prison: “Ha, I told you guys it would work!”

Let's check back on this story a year from now

People who invested in Thernos: Here we go again 😂

Saying Elizabeth Holmes was "ahead of her time" is like saying Star Trek invented warp drives.

So combine the fraud of Theranos with the hype of AI. What can go wrong?

When you knowingly make fraudulent claims that you have no possibility of fulfilling, that is not being just a little ahead of your time.

Theranos was an example of when I thought that if they had just properly used money invested in their research toward actual research, they probably would have achieved some technological marvels.

"... buzzword ... buzzword ... AI ... one drop of blood ..." lol

_"How many times do we have to teach you this lesson, old man?"_

The "drop of blood" idea is such a red herring. If they discovered a test for pancreatic or colon cancer that could run on a normal blood draw that would be a huge discovery.

Psychedelics are just an exceptional mental health breakthrough. It's quite fascinating how effective they are against depression and anxiety. Saved my life.

If you're outside the US, this test will be a nice affordable $5.00. If you're in the US, it'll be a nice unaffordable $18,525.

This sounds familiar...

Big Pharma about to charge $1000 for each AI diagnosis.

Fool me once, shame on you. Fool me, you can’t get fooled again.

I'm 40 and was diagnosed with cancer (DLBCL) in February and it wasn't anything to do with my lifestyle (healthy diet, didn't smoke/drink and exercised 4-5 times a week) but was merely unlucky due to a random mutation. The whole diagnosis was easily the worst part (even worse than chemo in some ways) so anything that can be done to speed up this process and save more lives will make such a difference.

One of the few things that machine learning is especially good at is recognizing simple but subtle patterns. Feed an enormous amount of data that you think contains a subtle pattern along with the outcome you think that pattern should detect, and it is exceptionally good at doing so. That's why it's much better to call this machine learning, and not AI. You are teaching a model to recognize a very specific thing, and for that it works well. That's one of the major differences between this and what Theranos was trying to do. Theranos wanted to build a general purpose blood test machine, doing dozens or even hundreds of tests on-site resulting in measurement, not probability. One of the important things to remember is that the part that Theranos couldn't solve is still unsolved here; making a cheap and compact machine to break down a blood sample in many different ways to get different readings. This, however, uses traditional methods of breaking down the blood, paired with a sophisticated model to compare those numbers to thousands of input samples. The ability of machine learning to identify patterns is not something that should be in question; and the people who need to fear their jobs are essentially those who have very little creative input to offer. If your job requires you to adapt an approach on the fly, you're still safe. If your job is to look at pictures and find a pattern, or collect very specific information from a controlled source, what you should fear isn't "AI", it's simple machine learning that can perform a similar process endlessly with hundreds and hundreds of years of practice to "learn" from.

I love how you addressed my first thought right in the second part in the title 😂

I think that the fact it’s been peer reviewed by John’s Hopkins makes this feel a little more legit.

The irony of using a petrol cars exhaust as an example when it’s that exact thing which contributes to cancer cases through air pollution

People in the comments are missing the point by immediately dismissing this because of what happened with Theranos. In the case of Theranos, it was a completely empty promise and they refused to share the scientific basis for their claims because, as we all know, there was none. This is literally the exact opposite situation. It’s not some Sketchy startup trying to sell investors; it’s actual researchers putting out peer reviewed studies that point to the feasibility of this. Informed skepticism is good. Boneheaded skepticism without actually looking at the science is bad

Cancer is misunderstood as a single but it is Category of disease in which there is abnormal cell growth which spread to other part of body. It could be the skin, brain, liver, prostate, or breast, among other organs. Every cancer is unique, and so are the treatments. If biomarkers are detected, the cancer is advanced and has a poor prognosis.

How many false positives? Yeah that's what I thought

Super skeptical of this one

all i want is my hair back

This is the type of uses I love to see with Ai, even if it’s in early stages it’s great to see people attempting to apply it in ways that can change people’s lives dramatically

I have met, in person, so many people with normal “biomarkers” who feel like they are dying; which makes me quite skeptical of this science.

of course we should be skeptical

In Mexico, the lab I visited advertised a CEA test to detect cancer - about 200 pesos or $10. If your score was over 5,. you had a problem and would need a complete cancer detection scan. When I was diagnosed, my CEA score was 15.8.

This is where I want AI.

But we all have cancer (our natural killer cells keep on top of cancerous cells if we are healthy) so surely this is pointless as it would just give a load of false positives? Surely it'd be better to monitor people's natural killer cell activity and people with low activity can take steps to improve their health before they get full blown cancer.

Gee... I wonder why early-onset cancer has increased by 1000 something percent. What could it be?

There is also the Galleri test by GRAIL, which is pending FDA approval in the US and has an accuracy of over 95% in detecting more than 50 different types of cancer using biomarkers.

Congratulations on your 500th Video 🎊💐

Lol the fact that no one could see that E.H was a psychopath is beyond me. You see it within 1 second.

My sister-in-law is suffering from liver cancer and this is great news that might spare others from having to go what she is enduring. Thanks for this report. Hope is wonderful thing.

This would be revolutionary, really hope this new way of detection will succeed! Amazing video, Dagogo!

It's surprising that you didn't even mention the Grail Gallery test. It can detect more than 50 cancers using the biomarkers and AI.

7:12 So they didn't treat that area and cancer came 2 years later? Why was it not investigated or treated beforehand?

From my physician’s point of view, cancer may seem more common because we can cure or control the older main causes of death, including cardiovascular and diabetes. On a positive note, we are now treating cancer better and better. In addition, earlier detection of breast cancer with modern digital mammography and better use of the prostate cancer test called PSA are helping us too these cancers much earlier and that managed them. Still, this tech looks quite promising. Thank you.🙏

Fyi: AI is an umbrella term. Some of the well known types: - Machine Learning (data learning) - Deep Learning (neural networks) - NLP (language understanding, includes GPT) This comment is written by ChatGPT-4o.

Interesting! Thanks for sharing it. ✌️

This isn't AI, just machine learning lol

I’ll believe it when I see it.

hope you do a follow up vid on this.

This is awesome! Finally we know exactly when it's necessary to apply thoughts and prayers!

Thanks for working so hard to put another video out for us 😊

I hate to break the fun and success of it but as a software engineer that is well informed and up to date to the current trend this might be another AI hype scams which I really wished is not.

Holmes was ahead of her time in a sense that she had no actual research to back up her claims. If Theranos started around now it could scam people much longer or possibly even fake it till it made it.

There are so many misdiagnosis in Psychology, I really wish one day through patient consent we'll be able to record the patient face, facial expression, body language and speech to have AI pattern recognition as an assistant in the diagnosis process. Especially the Brain Disorders gets misdiagnosed a lot particularly outside the anglosphere

"Ah sh*t here we go again"

Yeah, I'm not going to hold my breath. This sounds too good to be true, same as Theranos. If it works, good, I'll believe it when they can prove it

This is both fascinating and hopeful. AI's role in early cancer detection can't be overstated. It would truly be revolutionary if such technology could become universally accessible, cheap, and reliable.

Another big hurdle is to get treatments that can CURE it without taking a Machiavellian toll on the person's body, leaving damage and ironically a further risk of cancers, as well as draining their bank accounts. Many are choosing to skip treatment to avoid these two things. If we want to attack cancer with any degree of success, we need to address the whole picture.

ACTAVATE MEGA SKEPISEM.

Babe, Dagogo dropped another one 😊

The guy biting doctor's hand at the end was funny.

you need to release that outro track!

Amazing progress!

Theranos was a different thing though, it promised being able to do hundreds of different tests with one spoon of blood. This is just one thing.

Another well made top video

Ill wait for the peer reviewed papers on this.

People don’t seem to understand what AI really is. Not even AI people seem to understand what AI really is.

Yeah, I’m still going to be skeptical

My grandma had cancer, and then both of my parents had cancer at the same time when they were in their late 30's. My dad didn't make it. I'm now 30 and am fairly certain I'll get it too just based on how prominent it's been in my family. Though I can't afford the treatment at all, so I guess if I get it, that's it.

Where have I heard that in the past? Exactly, Theranos. Wait, what? AI to the help?

Unless this turns out to be a lie too. I thought you already did a video about how many of these AI companies are lying

Elizabeth suing from jail be like “This is my idea pay me a million dollars.” 😂

Can you please make video of you past videos as to how the idea and possibility you shared with us have unfold since Coldfustion TV

Regarding lung cancer in Asian women, this is due to them doing most of the cooking. Asian food is usually very oily and cooked at high temps, which creates a lot of fine particles in the air, which damages the lungs.

I work with cancer diagnostics. We are validating detection with plasma and blood at 10 ng input (equivalent to a drop of blood as lokd as their is enough DNA). We work with validating different genes across different organs. We already have FDA approval for tissue resections but ultimately the goals is liquid biopsy. The "problem" with this dry blood detection method is that it doesn't tell you what gene mutation you have, so you would not know what personalized treatment to give, and you would default to chemotherapy.

As someone with a family history of cancer, this is good news.

Oh, please, let this work. Getting cancer diagnosed is sometimes terrifyingly inefficient. Anything to help spot it sooner!!

i have never regretted subscribing to this channel

She wasn't ahead of her time, she was a con-artist. Now tech is catching up to sci-fi. Nothing more.

I agree . This is the type of application AI should be used for . I'd say in less than 5 years this will be used on a larger scale with how quickly tech moves .

Excellent.

Elizabeth Holmes v2

Did not even take 1 second before the Thernos comments came 💀

That anti theranos title got me XD

So the whole problem with using AI is that it will have the same limitations as any other diagnostic tool. You have to have a certain volume of a sample, in this case blood, to account for the probability of fluctuations within a give sample. For example, if I were to put 1,000 pieces of paper each inscribed with a letter from the alphabet, and had specific quantities of each of them, you could not properly predict (or even reliably come close to) the ratio of values for each at small sample sizes. If you were to pull out one piece of paper, you could not predict the mix. Pulling out 5 pieces of paper isn't all that more useful than one. For certain diseases a drop of blood may be indicative of the presence of disease, if the presence of something in the blood is entirely unexpected in healthy subjects...however you'd still have to do more testing after that. That being said, just because it did NOT test positive does not mean they are not sick, which is a bit problematic.

Honestly, FOR GOD SAKE, STOP CALLING IT AI. It's MACHINE LEARNING or DEEP LEARNING, NOT AI.

Just think about the potential treatments that exist in plants we have yet to discover. Scientists estimate that about 100,000 plants are yet to be discovered. Nearly all medicine treatments come from plants. This is another reason why it’s so important we protect our environment.

Elizabeth Holmes at parole hearing: Well, I told you it works.

I have knew it for a long time (years) now that we need AI if we want to leap forward, but in the end it could kill us. that being said we still need it to help us make real progress faster.

Turbo cancer is definitely not caused by the "Safe and Effective" (TM).

2:30 call me crazy, but what about COVID shots? The other things that you comented are not really that new

The old just aren't inflicted. The mother of my kids, and my best friend in this entire world passed away from cancer and she was just in her late-ish 30s. RIP Stace

Imagine if the world was in a crazy spiral after the discovery of a new mysterious illness and then suddenly, less than two years later, there was a new medical "treatment" that was not optional in some countries that would a few years later be skyrocketing cancer and heart problems. Imagine...

One of the main reasons theranos was able to scam so much money, was RU-vidrs like yourself making videos about science that hasnt even been verified at all yet, and hyping it up

China and AI makes me question accuracy lol

Perpetuating a fraud does not make one ahead of their time.

*My sister died 2 years ago of cervical cancer. She was only 39. My mom died 6 weeks later at 59.*

Dogs can do it just by smelling it. Blows my mind 100s of millions aren’t spend on training more dogs. They have like a 90% accuracy rate

Chinese and Ai what a great and truthful combination

I remember that the problem was that the single drop sample changed a lot. 2 samples taked 10 minutes apart would have very different results because has very little blood to average all the metabolics in the blood. A normal blood test has more reliable results because the blood was more mixed. I remember that this was the problem with theranos and not the technic used to analise.

This will also cause a lot of people excluded from insurance coverage due to the potential of developing cancer due to early detection by AI.

So can dogs