Theranos had other problems, like a fundamentally untenable design that many biologists criticized as being physically impossible. Not hard, but physically impossible.
Nevertheless it is true that moving fast increases risk. You either fail spectacularly or succeed spectacularly.
Some of the things Musk has claimed are probably physically impossible using the device Neuralink has constructed. For example, there is indeed a treatment for depression that involves stimulating parts of the brain using electrodes. But the parts you need to reach are way in there[1]. Neuralink's device is not capable of reaching these areas in its current form.
However I'm way out of my depth on this. The best I can do is listen to skepticism and try to evaluate it on its merits vs. the things that Musk says and his track record of over-promising.
I won't speak at all on Neuralink, but you don't necessarily need to stimulate that deep into the brain directly. You can piggyback off existing functional connectivity between that deep area and an area closer to the cortical surface. Functional connectivity is when one area sends activation or inhibitory signals to another area when stimulated. This is the mechanism behind using transcranial magnetic stimulation (TMS) to treat depression ([0] as just one example) despite TMS's shallow stimulation depth. Using the functional connectivity also provides an avenue of sending inhibitory signals, not just activation signals. With that said, it requires functional imaging (fMRI) to determine the functional connections. Functional connectivity also isn't indefinitely static and it does change over time. .
I should add that it's much more nuanced and complicated than how I've explained. Along with determining if a functional connection is excitatory or inhibitory, you need to ensure its direction as well because functional connections aren't necessarily bidirectional. You also want to be confident in how stimulating the target endpoint will affect the functional connection over time. It may potentially strengthen it (think "training") or weaken it (think "desensitizing"), thus potentially affecting the efficacy of the stimulation over time. As with most scientific fields, knowledge is built up one step at a time and you need to stand upon the shoulders of prior work before producing an application with confidence.
Once Theranos's patents lost their teeth, devices similar to what was in those patents appeared on the market within a few years. All from big pharma companies. Today, a lot of blood testing machines can run many tests on a few hundred microliters of blood.
When you go get blood work done, they collect a vial, but often use only a few drops at a time.
Usually not, actually, although some more complicated tests do involve sample prep. For most tests, it's a small microfluidic cartridge and a tiny amount of blood.
I would remain skeptical about this claim. There are a lot of exploratory tests that can be done with a drop of blood and have a solid history behind, for more complicated tests we would fall again in a "Theranos magic" case.
I hope that the investors finally learn their lesson and accepted that if you want to produce good results you need to starting with good data. In science there is not sense in finding the minimum algorithm possible that vomits the desired results. Alive beings are complicated
So either the claim that somebody is building the nbext theranos machine or recycling theranos machines for doing what theranos couldn't... either is false, or is just old technology wrapped in glitter, or would send serious doubts in the confidence in the results. It depends on the details.
or barely succeed and have to make up the debt to technical design and process quality later, taking longer in the end than if you had built a solid foundation for your endeavor
Nevertheless it is true that moving fast increases risk. You either fail spectacularly or succeed spectacularly.