autonomydriver

Joined 7 years ago

Karma
330
Submissions
54
Member since
Nov 16
View on Hacker News

Recent submissions

Comment · 11 months ago
Poor wording, should mean that each finger has a dedicated output, but you are right, they are all tied together
Comment · 11 months ago
In my searches, yes, I found something like that from an italian university. Sadly can't find the paper now.
Comment · 11 months ago
Coming in a V2, semester just started again, so I have a little less time for it. Given that I also have some other projects on the way :)
Comment · 11 months ago
The two BJTs do limit current to 20mA on all finger outputs. Voltage feedback is also set up to ensure maximum 32V. MCU protection, well just a ferrite bead.
Comment · 2 years ago
Especially interesting for neuromorphic computing hardware. Has already been shown by IBM's TrueNorth and NorthPole chips. Tough in my opinion, a major bottleneck, is the lack of bandwidth on these chips. They can be highly parallel for their circuits, though for example image data, only 28x28 pixels, would already need 784 DACs coming from a digital source.
Comment · 5 years ago
Often times in my opinion, those people then when they get released are a whole different person and you can learn lots from them about life and especically appreciating things we (outside) take for granted. And in my opinion such a program solves the last step after the release, and that is reintegrating.
Comment · 5 years ago
99% of SPACs are scam. They really are, because if they weren't going public via an IPO there must be something wrong and if not they had reasons why not to go public, smarter for companies working on government projects etc.
Comment · 5 years ago
Well done you will deliver faster than all the other companies with the regulatory overhead.
Comment · 5 years ago
Show's again how blind universities are with millions of funding, a stundent has to show them.
Comment · 5 years ago
This is a very cool blog, thanks for sharing, had no problems opening the page in chrome on windows.
Comment · 6 years ago
I tried it in the city center it was going nowhere, the bigger and complex the models get the more they learn "intuitively" i.e. avoiding pedestrians, avoiding crashes etc. Like the video where Autopilot predicted a crash 3 seconds before it actually occured.
Comment · 6 years ago
Thanks, yes maybe I will add proper sensoring to the project with emulating a can bus.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton