I'll grant that we could send humans to mars sooner if we really wanted to. My point is that not achieving a bigger dream doesn't make current progress a hype wave followed by a winter.
And "LLM's just make plausible looking but incorrect text" is silly when that text is more correct than the average adult a large percentage of the time.
this attitude is so ridiculously disingenuous. Surely if a computer can score incredibly well on math olympiad questions, among other things, "a computer can make plausible-looking but incorrect sentences" is dismissive at best.
I have no idea about AGI but honestly how can you use claude or chatgpt and come away unimpressed? It's like looking at spaceX and saying golly the space winter is going to be harsh because they haven't gotten to Mars yet.
they didn’t submit it here - I’m not sure they bear a responsibility to make sure their work is accessible to HN or any other target audience. If you don’t find an otherwise respectful and substantive blog post useful, perhaps best to leave it be.
An unsophisticated person looks at this and says golly I better get tan!
A somewhat sophisticated person says something like the above.
People who change the world will wonder there’s something there. especially given the sober and candid examination of possible confounding factors and mechanisms of action.
“There are approximately seven scanners per million inhabitants and over 90% are concentrated in high-income countries. We describe an ultra-low-field brain MRI scanner that operates using a standard AC power outlet and is low cost to build.”
This is fantastic. What a sentence to get to write.
> There are few good passive investments right now. There's so much capital available in search of returns that all the good ones and most of the mediocre ones have been taken.
This is one of the more insightful comments in this thread. The risk is not merely that Sam’s investments are not diversified, it’s that Sam has competition from more sophisticated capital. Truly passive investments often scale (and get more passive with scale). That scale attracts a lot of capital that can afford lower revenue, higher risk, and better expert knowledge.
I'm not sure that's necessarily true. It is true (at least with a non-constructive proof) that if you pick a 'random' real number then it contains all possible PDFs with probability one ( or that the set of numbers for which this is not true has lebesgue measure zero). But I'm not sure it's known that pi has this property.
Surely there's something to learn here though. I haven't read the original paper but a quantity that's preserved across brain scales is either an artifact or a neat insight.
Your criticism reads like someone accusing economists of being outrageously misleading when they don't sample individual households but measure macro indicators. It's like saying Ramon y cajal was ridiculous because he couldn't image the neuropil effectively. Or like saying early optogenetics experiments were ridiculous because who knows if you're stimulating a neuron in a realistic manner?
And in any case, it's true that synapses are comically small relative to voxel size, but we also have some reasonable information about projection patterns and synapse number from various tracer or rabies studies with which you are no doubt familiar.
I haven't read the nature paper the press release is about and I'm not a huge fan of many d/fMRI practices or derived claims. And I've worked with enough mammalian dwi data to be skeptical of specific connection claims. But this strikes me as a rather interesting result even if you can't measure all the synapses at the right resolution: either the tractography method has connectivity conservation artifacts baked in, or there's something interesting going on.
In my experience, you are way off base. Doctors can be very arrogant but I've never met someone from another profession who could point me to research articles regarding their proposed plan of action. Doctors at major hospitals are often either a) residents who are in their nth year of leaning post med school, or b) expected to publish at least case study papers regularly or communicate with those that do.
Your description of the localization approach is accurate, but this system is almost certainly Brainsight, given the screen shown in the image at the bottom, the IR camera in that image, and the shape of the tracker arm.
The general approach is straightforward: MRIs are have real-world coordinates. Anything on the head (TMS, EEG, a surgical instrument) also has real-world coordinates. To co-register the two, you need to associate 1) markers at MRI time 2) markers at TMS time.
Once you have that correspondence, you can position any other objects relative to either one, like surgical instruments with reflective markers or TMS systems or whatever.