Hacker News new | comments | show | ask | jobs | submit login
Kernel, a startup working on brain implants to link thoughts to computers (technologyreview.com)
50 points by t23 310 days ago | hide | past | web | favorite | 11 comments

I do some of my Master's research in this space -- specifically trying to build computer programs to interpret what a person is thinking of based on the neural activity occurring within the brain. This field is so, so far away from any sort of useful brain-computer interface, but there is exciting work going on right now.

The biggest problem is that we don't understand how the brain works well enough. It's quite complex and pulling out useful information is very challenging. Machine learning algorithms can be ran on neural data (fMRI/EEG/MEG) to identify some simple words or detect specific intentions/reactions to stimuli, but we definitely aren't getting this information at a rate which is anywhere near faster than current human interaction methods and with nowhere near the accuracy (in real world settings) we would need to be practically useful.

Pardon my relative ignorance on the subject, but couldn't we take an approach in which we attempt to let the brain figure out how to interpret/send signals? I.e. find a way to take advantage of the well-established extreme plasticity of the brain, and let it figure out our interface, rather than the other way around?

There are some cases where plasticity can be useful. For example: if a particular intention is associated with the neural activity while a person lifts their arm, eventually the person can train themselves to trigger the intention just by thinking of lifting their arm (without actually doing it). This is the brain learning to control the intention instead of the action.

I seriously doubt that we'll ever get to the point that we're comfortable with brain surgery just to upgrade our "connectedness." The real win is some sort of skull cap which can read your brain signals, and communicate with your brain through the skull. Seems impossible now, but I wouldn't be surprised if they come up with a way.

My concern is that we haven't yet gotten smartphones right. Most of us love and hate them. I'd love to have something like parental controls on my phone which I could use to limit what I can do with it: set limits on my usage, control which apps I can use, and limit what I can do with them. There's a variety of enforcement mechanisms ranging from putting in a longer random password and not writing it down, have my spouse set it up, have it setup initially and then come up for review every three months...

But imagine if we all had something similar, inescapable, engaging, entertaining, and utterly addicting in our heads. Yikes. Imagine if you've got notification overload...In your head. Yikes.

No thanks. I'm still trying to get a handle on my damn smartphone addiction.

This is worrying. The surveillance capabilities here alone are terrifying, not to mention the potential future invasive adverts we could be presented with. I hope this doesn't happen in my lifetime.

Not to worry. The implant will help you adjust to the idea. You'll be telling all your friends how great it is!

Lets hope the FSF make some progress with the larger public before we descend into such a hellscape.

While the whole brain thought interface is a great idea, I really just wish I had some kind of keyboard and mouse I could control with my mind, and a pair of glasses with my screens. That would eliminate so many ergonomic issues, and probably increase speed as well. I wonder what typing looks like in the brain...

We have a very basic understanding how our brains function. They are most complex bilogical thing we have ever encountered and due to morality issues it's very difficult to test hypothesises on. Sadly this man's quest will not come to fruition but a better allocation of effort would be to the get the NIH/WHO data set of scans be freely available for AI to start figuring out patterns. There wouldn't be any real economic activity to do so(the shame!) But what I believe is necessary to move the field forward.

I think a brain interface to computers and phones would be useful but privacy would really worry me. Hack someone's brain, and you get way too personal info. Sometimes when I'm reading on my tablet, I think it'd be cool if it would just scroll when I get towards the bottom. Eye tracking might work, but I feel like you'd falsely trigger it sometimes.

Also, I feel like I think way faster than I can type. I can visualize the piece of code in my head before I even type it out. Imagine no more spelling mistakes :)

Brain surgery already seems very scary, as you have to stay awake during it. Then when they come out with new hardware generations, go under the knife again? Brain surgeries also are expensive. But I could see how something like this would make me 100x more productive than I am now. So if I created something really cool using my new special abilities it would probably pay for itself.

So I see a lot of upsides and downsides. I don't know if this would be something I'd want to be an early adaptor of, though. Imagine if the company supporting it went out of business, now you have this unmaintained piece of hardware in your head. If something like this was perfected, I feel like this would be an asset to have just for the productivity advantages.

Well, speech provides 10-100 bps, which isn't much. But I doubt that pulling speech-driving data would do any better. You'd need to work at lower levels. There's probably lots of multidimensional parallel processing at play.

Requisite plug for Echopraxia by Peter Watts :)

Applications are open for YC Summer 2018

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact