Ask PG: Hardware/Infrastructure startups and YC? | Hacker News Reader

Ask PG: Hardware/Infrastructure startups and YC?

6 pointsmoxiemk15 comments

A number of the ideas I've had for new disruptive companies aren't traditional web applications or services, but instead are hardware producers, new takes on systems software, and re-imagining infrastructure. More SpaceX and NeXT than Hipmonk and Posterous.

These markets are full of slow-moving entrenched players: a ripe target for disruption in a major way. However, it's probably a harder market to get into, with more capital outlay required to get started.

My question is: would YCombinator be interested in such a startup? I imagine that the level of funding YC gives would help a team of founders get a purified idea of what they intend to make, and maybe some very early stage prototypes of software, though probably not hardware. Being successful would entirely depend on getting a funding round once we produce the plans so we can hire talent, purchase materials and start manufacturing.

View on HN

Startup Ideas We'd Like to Fund http://ycombinator.com/ideas.html

"27. Hardware/software hybrids. Most hackers find hardware projects alarming. You have to deal with messy, expensive physical stuff. But Meraki shows what you can do if you're willing to venture even a little way into hardware. There's a lot of low-hanging fruit in hardware; you can often do dramatically new things by making comparatively small tweaks to existing stuff.

Hardware is already mostly software. What I mean by a hardware/software hybrid is one in which software plays a very visible role. If you work on an idea of this type you'll tend to have the field to yourself, because most hackers are afraid of hardware, and most hardware companies can't write good software. (One reason your iPod isn't made by Sony is that Sony can't write iTunes.)"

Wow, as a hardware/software guy this is comforting to hear.

And it's true, there's lots of poorly architected software out in engineering-land.

I had been reading this list; the main reason I still have a question is because this seems to indicate a scale different than I think we can achieve.

Not so much the next Flip Camera company, taking a market that major companies think is a toy. Instead, going head-to-head on their products that they've allowed to become crufty square-holes for your round peg and making great tools for real problems, not adapting commodity PCs in ludicrous ways.

I'm working on software/tools for embedded field as well as trying to develop a wireless driver for 802.11p (automotive "WiFi").

It's actually not that expensive to get a prototype done with many of the hardware resource available for either free (as samples) or dirt cheap in small quantities.

Say you want to build a new Gaming Router, you can buy the underlying hardware to do development on for under $200 in quantities of < 10 and that will give you all the needed parts (board, radio, antennas, cables, enclosure, etc.) with the right specs to do just about anything you want. You'd then run Linux on it (OpenWRT is easy, comparatively) and wright the necessary software on top of that.

As the Ideas point #27, linked to by ig1, shows, most of the problem solving is in software.

Give you an example, I worked for a startup, 3 software guys (myself included) built a router that had all the capabilities of a Cisco small business-class router in 6 months from getting the hardware to being feature complete. No hardware people were involved. We used a crappy, 3 year-old platform that would over-heat (the SoC would start to literally smoke) if we ran more than 2 IPSec connections, but we were able to tune the kernel and our software to make it a viable product for a lot of businesses.

If you know your way around Linux, building embedded products is unbelievably simple.

If you or anyone else is interested in talking more about this field, shoot me an email (it's in my profile).

GearBox was a TechStars company from this summer that built a robotic ball: http://www.techstars.org/companies/orbotix/

Reply on news.ycombinator.com