Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
dopeboy
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
35 ms
·
331.
▲
by
dopeboy
6y ago
This is an incredible video - thanks for sharing. Couple notes for those like me who are new to the speed running world: * The goal here is to get through a particular level as fast as possible. Not to kill enemies or get points; it's
332.
▲
by
dopeboy
6y ago
> I hate that Chromium’s snap takes more than 10 seconds to load on cold boot on a freaking SSD, whereas .deb and Flatpak apps load in 1-2 seconds. Can someone verify this? As someone who will eventually upgrade to 20.04, this is concern
333.
▲
by
dopeboy
7y ago
I don't know if there is an equivalent. The beauty of conferences are the unscripted, chance encounters with other attendees. I gave a talk at last year's PyBay (bay area python conference) and the magic of it was in the side conv
334.
▲
by
dopeboy
7y ago
Brandon and Ishmael are awesome people, y'all. I remember sitting down with them in the early days; excited to see all the progress made since.
335.
▲
by
dopeboy
7y ago
With COVID-19 here, I imagine in person training has been affected. Can you talk through some of possible solutions at hand for the environment we're in?
336.
▲
by
dopeboy
7y ago
No need to go that deep to experience it. Dive 20+ meters and you'll feel it.
337.
▲
by
dopeboy
7y ago
Yep, citizen effort here as well in Oakland. Me and a couple folks created https://www.lanelookout.org to do precisely this. We got a lot of support from our local code for America chapter, OpenOakland.
338.
▲
Why I'm Working on a Startup
(dopeboy.github.io)
2 points
by
dopeboy
7y ago
|
0 comments
339.
▲
by
dopeboy
7y ago
I own an ebike (Ancheer) and a escooter (Xiaomi) as well; cool to see your analysis here. You're dead right about ability to handle bumps. With my scooter, I will watch the road in front of me very carefully to scan for rough areas; th
340.
▲
Why did we stop playing?
(dopeboy.github.io)
1 points
by
dopeboy
7y ago
|
0 comments
341.
▲
by
dopeboy
7y ago
Awesome questions & comments thomzi12. 1) Possibly. It's not something we've explored though I am told by many pet owners that getting them to stay still is a challenging task. 2) I might disagree with you there; we're tr
342.
▲
by
dopeboy
7y ago
hanniabu, this is an excellent suggestion. We were thinking about a chin rest (like what you see at an optometrist) but we didn't like how constricting it would be. I do like the idea of a headrest embedded with some sensors, though.
343.
▲
by
dopeboy
7y ago
Scandy ( https://www.scandy.co/ ) for the scanning, Meshlab for visualizing and computing the traces.
344.
▲
by
dopeboy
7y ago
It doesn't matter if you move after the scan, so long as the orientation of your head is the same, you're good. You do need to stay pretty still during the cut and we have the screen + camera up front to assist you with that.
345.
▲
by
dopeboy
7y ago
eplanit, right there with you. There's a couple variables ahead in our list (precision, speed, repeatability) but cleaning up the hair will eventually get added to that list. My hunch is we'll start exerting tighter control over t
346.
▲
by
dopeboy
7y ago
Great point rajeshshamara. Here's how we're defining the time involved in getting a haircut: (1) Time spent making an appointment (2) Time spent going to the barber (3) Time spent waiting the while barber is finishing their existi
347.
▲
by
dopeboy
7y ago
Awesome questions, I appreciate the curiosity. We're using the stock force control software and plan to tweak it soon. I couldn't tell you how it works under the hood. We use mannequin to test and then, other body parts [0]. Long
348.
▲
by
dopeboy
7y ago
Agreed. That's our system that keeps track of whether you stay still. It's a little too sensitive and needs some tweaking.
349.
▲
by
dopeboy
7y ago
This is the not the 'e' edition. I think the 3e's cost around ~$24k. The standard UR3 is discontinued as far as I know.
350.
▲
by
dopeboy
7y ago
I see your smile but I thought I'd answer anyways :) A while. We've had folks ask about it and we're going to stay away from it for now for the obvious reasons. Close to 50% of all haircuts are clipper cuts so we think there&
351.
▲
by
dopeboy
7y ago
Hi ASalazarMX - no it couldn't. We have software constraints around speed and position. Additionally, there's hardware checks that shut down the arm if it moves too fast and/or encounters too much pressure.
352.
▲
by
dopeboy
7y ago
Hey HN I'm the founder behind this. I'm a guy and I'm tired of: * Waiting for a haircut when I book an appointment. * The inconvenience of only having one barber that knows our style, especially when I'm traveling for wo
353.
▲
Show HN: I'm building a robot that performs haircuts [video]
(youtube.com)
186 points
by
dopeboy
7y ago
|
190 comments
354.
▲
by
dopeboy
7y ago
Ten years ago, in grad school, I co-wrote a video conferencing module in VHDL [0]. I haven't touched HDLs since but here's what I remember very clearly from that project: * It took > 3 minutes to compile our code. * DMA made a
355.
▲
by
dopeboy
7y ago
I've been to three of these and would highly recommend attending. It's a great way to network with others and also get a taste of the YC mindset.
356.
▲
by
dopeboy
7y ago
Sorry in advance. Thanks for offering this to us.
357.
▲
by
dopeboy
7y ago
I segmented the beginner crowd into two cohorts: side project folk and operations people ducktaping zapier+airtable, etc. The former does not have budget, the latter does (small businesses, early stage startups).
358.
▲
by
dopeboy
7y ago
I'd be easy on them. They're still in their early days and want to exercise control over who receives the product. They probably have a good idea about who it's a fit for and who it isn't. Additionally, they probably wan
359.
▲
by
dopeboy
7y ago
I had a chance to sit down with Ellen, one of the co-founders, and see an earlier beta. Congratulations to the entire team on this milestone; excited to see what you have in store. For context, I used to run a consulting company where I shi
360.
▲
by
dopeboy
7y ago
We're still figuring it out is the honest answer. Here are some learnings from this past year that address some of your points: * Robotic arms are getting better. They're becoming easy to interface with (I programmed the arm in th
More ›