Hacker News new | past | comments | ask | show | jobs | submit login
How pCell works and why its a bigger deal than anyone realizes (akbars.net)
281 points by followben on Feb 28, 2014 | hide | past | web | favorite | 114 comments

A couple of clarifications and comments (new whitepaper is coming with a lot more info):

The iPhones (and dongles) we demoed have no software modifications at all, and there is no additional overhead beyond the standard LTE protocol. pCell works with unmodified Android LTE phones, too. (We had wanted to post a lab demo we did of the Galaxy S4 and an Xperia, but we just ran out of time.) Out-of-the-box compatibility is essential for rapid deployment. And, yes, we are meeting many of the core goals of 5G today (e.g. unlimited 4K UltraHD streaming) using Rel. 8 LTE devices in LTE spectrum.

pCell is indeed protocol agnostic, and can concurrently support different protocols in different pCells in the same spectrum. For example, we can support unmodified LTE phones in their own pCells, while concurrently supporting lower cost/lower power devices with far lighter protocols (that are lower latency), since we don't need all of the complexity of LTE. For example, there are no cells, no cell edges (and no need for CoMP), and no cell handoff.

I'm waiting for something to connect the dots and realize this is a vastly more efficient way to use white spaces than anything currently on the table. White spaces will be full overnight with current techniques. With pCell, they will never be full.

More data is coming later. Apologies. We have been utterly overwhelmed in incoming inquiries since launch.

But, I will confirm this: pCell is indeed a much bigger deal than anyone has yet touched on. The "tubes to transistors" analogy is not just marketing speak: Compared to cellular, pCell is far more reliable, enables much smaller and lower power device and can be continually extended in density. Tubes had physical constraints that limited their reliability and scalability. Transistors did not. Cellular (and other interference avoidance protocols like Wi-Fi and cognitive radio) have a physical constraints that limit their reliability and scalability. pCell does not (as far as we know). Cellular has stalled in scalability. There is an entire era of innovation in front of us with pCell. - Steve Perlman

Thanks for the clarifications but it's still hard to get it.

Is pCell an alias of Picocell?

What's the difference between a pCell and a Femtocell with cooperative MIMO and distributed antenna system (DAS)?

"pCell is indeed protocol agnostic" means nothing has been modified in LTE Rx/Tx physical layer or data link layer right? Then does that mean all the inventions are in the antenna system?

Without cellular, how do you handle the dead zone issue, because real time tracking & beaming will be blocked by walls?

I think the p stands for "personal" ie. it's so tiny that you can create one for each device.

From a technical point of view, it would be interesting to know how this performs near an airport. Multiple giant RF-reflective surfaces travelling much faster than 70mph must be very hard to deal with. Perhaps the lower per-transmitter power makes reflections at altitude less of an issue?

I've experimented with similar things in the past, and mainaining coherence for long enough outside the lab is Very Hard, so congrats on doing this.

Best thing since sliced bread? If Anything this fundamentally solve the Mobile TV problem. or Heck all Internet Connection problem with one go. No more silly ADSL Phone line or Cable Modem which are far worst then today's LTE. It would get rid of Home Router as well, as long as all of our Mobile Devices have LTE. I cant wait to pay for it.

And it seems we move the bottleneck to the cell backbone. Which could get saturated easily.

Until a mobile connection will reach the speed and stability of my 100Mbps cable connection some time will pass i guess. It also would need to be unmetered as right now the max i can get for LTE in Germany is 30-50GB for home connections. I know its just a matter of when, but still...

It properly still wouldn't be unmetered ( Or it could be in less dense population ). But the Cap could be much higher then what we are getting now. The reason why we are all capped at this moment is exactly what pCell is trying to solve. Sharing of Spectrum. The current LTE could already do 300Mbps, and with the accuracy of pCell there is no reason not to believe you cant get faster then 100Mbps in real world. Which is like i said, you dont really need your Wired Internet Line anymore.

> "no cell handoff" I'm curious how this would work. Presumably in the pCell system the base station focuses the radio signals in something like a beam towards the device but this must have a finite range - I wonder how far - and when you go out of range the system must switch to another base station. Maybe you call this a different name than 'handoff' but presumably the switching base stations thing must happen some how?

The pCell is "synthesized" using all base stations in range I guess. Which base stations that are involved in your pCell changes as you move around ...

Ah - I see, thanks. I guess that's how they can focus the signal so narrowly to 1 cm or so. Looking at the numbers the wavelengths used for 4G are of the order of 1 ft so you wouldn't get much focus from a single aerial of similar dimensions.

Since multiple base stations are involved, there is handoff where the far away ones leave and new ones join - but unlike cellular systems, the end device doesn't have to do anything during the handoff, it's transparent.

Thanks for the info! Seems like a neat system.

Really curious to hear how it's focusing into a 1 cm ball frequencies for which 1/4 wave is significantly larger than that.

Which leads me to wonder about it's possible application to audio. Headphone free ideal headphones with everybody in the room hearing a different source.

Will this technique work for license free systems like WIFI ?

802.11n already does this with MIMO

Answered in the article.

I hope you can eventually find a way to work onlive (or its successor) into this (I imagine this is one of your reasons for approaching this problem?).

I think the idea of focusing multiple low power radios onto a single point so that they coalesce is simply being rediscovered. While we could do this before the real ingenuity is being able to rapidly update all the parameters when the point you want to focus onto is rapidly changing.

Another scary idea: While we could beam wireless power imagine if this was in any way weaponized. With all these stations beaming power to everyone what would it take for an overwhelming number of them to direct it at a target with the purpose of eradicating it? Certainly not a laser in the traditional sense but surely something just as destructive.

The idea of deliberately burning out devices is interesting, and probably feasible, given a dense enough array of powerful enough transmitters (bear in mind that one of the key benefits of pCell is that it uses far lower-powered transmitters). But it seems to me the inverse square law is going to make it a very inefficient weapon.

Inverse square law does not apply to this as usual (although it does if you fix the beam configuration) -- but you can take your tin foil hats off (I guess this would indeed make a good scenario for tin foil hats) since you need real time channel estimation and feedback.

Are you sure the inverse square law does not apply? I think you're just getting a bigger coefficient. You might be constructively interfering signals, but the signals are still diminishing over distance. It's not magic, or they wouldn't need to place the antennas all over the place.

It doesn't have to be high power to be effective. Google search on "russian microwave US embassy" (not Bing search, unfortunately), which yields some interesting background on the Russian's beaming microwaves at the US Embassy. It was within the US legal limit at the time. Health problem ensued. Making an adversary sick without noticing may be more effective for some situations.


The literature seems to indicate that the Russian microwave campaign resulted in no measurable health effect to the employees at the embassy. This is unsurprising because a wide variety of publications detailing similar microwave exposure from modern electronic equipment like cellular towers, phones, and WiFi access points seem to also indicate a negligible widespread health effect.

If you can transmit power to a 1cm bubble in space, you may be able to offset it into, say, brain tissue. Brain tissue probably doesn't due too well under a focused heat source.

That's were magnetos helmet comes in. In all seriousness what material (or device) could stop this from working?

A simple Faraday cage, I would think. Hard part is making it look good.

we're going to need that transparent aluminum

aka sapphire & ruby

Actually I seem to recall that the eyes are much more sensitive to heat; the eyes have sensitive tissue and cannot shed heat as fast as the brain.

I really don't think the goal was to make anyone sick.

The Soviets were likely using microwave energy to activate passive listening devices hidden in the US embassy during its construction. ISTR the whole embassy had to be torn down and rebuilt.

Thanks to Snowden, today we know that the NSA refers to this technique as "flooding" and "radar" http://leaksource.info/2013/12/30/nsas-ant-division-catalog-...

The whole "make people sick" thing was likely a cover story to conceal the technique, even though both the US and the Soviets were clearly aware of it.

See also https://en.wikipedia.org/wiki/Thing_(listening_device)

Nice explanation, does it bother anyone else that you could set up a bunch of antennas such that they could create a lethal 1 cm ball of energy anywhere in the space they can see? Sure they think about powering devices but what about a couple of Watt-Seconds of energy appearing in your frontal lobes? If Perlman can do that its kind of the ultimate 'border' fence is it not?

(I'm tempted to write, "All you would need is an unscrupulous engineer to develop a targeting system ...")

First of all, your body would likely screw up the interference patterns enough that it wouldn't be a "ball of energy" in the center of the person - it would likely mostly diffusely heat them.

Second, the amount of power to burn a person is many orders of magnitude higher than needed to receive a signal. The transmitters would have to be extraordinarily powerful. I can only see this happening if they plan on making a wireless power transmission service.

Also, a 1cm ball is extremely optimistic, given that LTE frequences have several cm wavelengths. It would also require an enormous number of pCell transmitters to concentrate their transmission into such an area. For communications though, even concentrating into a 20m ball would be a huge improvement over existing technology, so I don't know if there is incentive to place more antennas than required for that.

> Second, the amount of power to burn a person is many orders of magnitude higher than needed to receive a signal.

And how does it compare to the power needed to make a car run? Did you look at the patents in the end?

Have no doubt that we'll increase the frequency untill we can focus it in a few cm sphere. The market will demand that.

I just replied to another thread with this exact same thought, then looked down to see yours. And yes, it is scary... If I understand it correctly, the targeting system just consists of normal receivers hacked to offset their targeting feedback by a distance that would park the energy ball inside the victim.

In any of the applications where this is deployed broadly, that is, where we cover our cities in transmitters, then we'll have effectively given over this remote kill capability to anyone who can do trilateration.

You would basically already do that with mobiles. When people use for talking they keep them <1cm from their head. And while 70mph is indeed great speed people sometimes take high speed trains (200mph) or don't turn off their mobile on an airplane (~300mph in vicinity of ground)

Tin foil hat? (Ok, maybe a Faraday cage hat.)

If this kind of technology were to fall into the wrong hands would we even know? Would it appear as though the victims had a heart attack or stroke?

You would know, if you did this to someones brain it would have lesions and damage. Although you would have to give them an MRI to find it. It would appear as if they had died of a stroke.

The first question that comes to mind is this: how can we defend ourselves from being brain burnt?

Funniest part about that is that a foil hat might very well prevent it from forming in the brain.

That is the funny part, and it would work. If you wore conductive clothing you could create a faraday cage around yourself thus preventing RF energy from crossing that boundary.

On a related note, discussing this offline with some friends they pointed out that the pCell system is getting feedback from the phone about its signal strength, and your body wouldn't be giving that feedback, so you would have to have something tied into the pCell on your person (or not move). Your phone could work for that. The question then is how far away from the phone can the energy be localized while the feedback from the phone would still be accurate enough. That I don't know.

Yea that's one thing I'm not sure about is how easily one could create it without that feedback. I'm sure it'd be possible but certainly more difficult.

You can already cause injuries with phased array radars, right?

Yes (per the guidelines the Navy issues on working around the Aegis system), but not quite like this. Per the discussion about phased array in the article the amount of power you need to do that is many times the amount of powered that gets delivered. This technique allows you to create such an interference node without any excess power from any single source.

What seems to me a more realistic concern is that you could create a beam of energy that would be lethal to an individual device. Brains are typically a lot more durable to EM radiation than electronics.

And more generally, it seems like relying on coordinating devices you don't physically own opens you up to all sorts of new kinds of DoS attack.

Actually if they could make it that small wouldn't it also be useful for surgery? Imagine non invasive surgery.

If the latency of pCell is low enough then you could have a thin-client phone tunneling into a cloud OS. Now get rid of most of the CPU/memory hardware complexity on the phone (which should lower power requirements for the phone too). If you can wirelessly send energy then get rid of most of the battery.

Suddenly you've got a very, very thin phone with infinite battery life and computational power only limited by the server you're hooked up to. It's the perfect mobile device. Do the same thing for a laptop and I'm in paradise.

Wow, what a paradise. Total control made simple, that's what I see...

Device good enough to do encryption?

It's not like you can just sprinkle some encription there and it will be ok.

Self updating newspapers, who needs history with an economic engine like that!

Sorry to rain on your parade, but don't you like idea that computing device is independent from others and does not require network connectivity 100% of time when you want something done?

Depends on the device. How about this:

An extremely thin tablet that's essentially a touch screen connected to your home computer, aimed for use in the home. Tiny battery, very little computational power & storage, hooked up to your own beefy computer.

Edit - intended for use just in the home. I'd want one, certainly.

Too bad screens are using a good chunk of your device's battery. Reducing the radio and computation to close to zero may reduce consumption to 30-50% but it's not an order of magnitude. The batteries are here to stay I'm afraid - until we get some radically different display technology, along the lines of a color, 30 hertz e-ink.

This was more about having wireless power too (part of the parent comment). You don't need a 10 hour battery on something if you can power it wirelessly.

if pCell also enables wireless power, youd just need small buffer battery.

Well maybe sometimes, but given all the advantages of the envisaged connected approach, I'm not sure any more if the advantages of going off line will outweigh the disadvantages.

Do yourself a favor and get of town more. Leave the phone behind!

? I'm not sure I understand your point - I go out of town plenty, and I leave phone and computers behind. When I need phones or computers, I'm generally in or near urban areas.

And then all you'd need to do is carry a briefcase battery.

Assuming pCell works as advertised it's a huge deal.

It allows you to cover a city in cells for insanely cheaper than current cell technology and provide far, far better reception and bandwidth to far larger numbers of customers AND it works with existing LTE gear while also affording simpler, cheaper, and lighter-weight new gear and reducing power consumption across the board. What's not to like?

I don't think anyone who has heard about pCell thinks it isn't a huge deal, if it works.

One example of a major potential problem: how well does it track erratically moving objects (e.g. most cell phones?), in built-up areas with lots of signal reflection?

The answer would largely depend on how large the focus area is or how fast the signal quality degrades the further from the epicenter of the focal point that the receiver is.

Maybe the system could apply road map info to fast-moving cell phones which it can assume are in traffic. And this could help it predict possible future locations of your device.

For much of the day the majority of devices in a given "cell" will be semi-immobile at a desk or slowly moving when someone's walking.

What happens when someone turns on their phone on a plane? ;-P

I think an on-board "repeater" would be the best solution. Of course, a high-powered directional antenna would translate to less hopping between towers but also increased costs to the carriers. Something like Google's Loon project might have more coverage and work well with aircraft.

> in built-up areas with lots of signal reflection?

If there are multiple plausible positions due to multipath, it probably wouldn’t hurt to create multiple cells for one device.

Steve Perlman always comes across as a snake oil salesman, but I didn't think he could do OnLive either, and that technology works fantastically. As a result, I'm pretty excited about pCell, even though it's being preceded by a wave of marketing hype.

It still hasn't made its way to my country, but it certainly is an interesting service.


However, they did overpromise and under-deliver, but it's still very impressive.

As many positive things as I have heard about it I have one question: will we ever actually see this as an actual product? I feel like I hear about so many cool things which never come to fruition and I have a nagging feeling that this might be one of them. Why? Because, as I've read, the companies would need data centers to manage all of these connections. What interest do these companies have to risk all of this money (a recurring cost, might I add) when people probably wouldnt spend more for it? I dont profess to know, so can someone who has actual information enlighten me?

That isn't why stuff doesn't turn out. It usually has some problem or another which is being assumed will be "solvable" by the time they release.

Never solve it? It's never released.

Alternatively, it does come out, but its gains are just evolutionary improvements not revolutionary.

"With this technology, it's conceivable that 5G wireless could displace both cable and DSL connections within a few years, as is claimed in the presentation."

A few years?! This statement should give landline data providers, like Cox and Comcast, some serious pause.

Except for the part where you need an huge array of transmitters with backbone bandwidth equal or greater the number of devices they may be serving simultaneously.

In which case those companies just start doing fiber installs to wireless stations, not houses, at roughly the same density.

Or just piggyback an antenna into the hardware otherwise installed at each consumer location. There was a theory almost a decade ago that Google would do something like that on top of all the 'dark fiber' it was buying up at the time.

Why, their wireless counterparts will surely buy spectrum or what not to make happen during a 20yr adoption cycle in the US.

The original Columbia presentation is actually pretty cool ( https://news.ycombinator.com/item?id=7274288 ) I thought I'd mention it as it only got a couple of votes and sarky comments at the time but having just watched it all I think it really look like a major breakthrough, maybe the biggest one of the year and could be giving us 4k video all over the place before long. One of the interesting things is it will work with existing unlicensed spectrum, existing handsets and the backhaul can be done with line of site links on rooftops so it can be rolled out pretty much straight off - well they are talking a year or so.

Once you begin exploring the implications and the possible applications to things other than communications it begins to look like possibly the biggest breakthrough in history.

Only one in energy could surpass it.

I wonder if wireless power like this could be used to provide power for launching spacecraft and this avoiding the tyranny of the rocket equation?

This wouldn't help a rocket. For a rocket you should just aim a laser at it.

But you still need reaction mass, and pure energy doesn't work well for that.

Space-based applications could include a network of base stations transmitting power to a cloud of worker drones.

Bio-based applications could include a means to power or recharge implanted devices and machines.

> Will it work for WiFi as well? It's protocol agnostic, so it could work in unlicensed spectrum as well. The issue is that you don't have complete control over all the other transmitters, so you can't coordinate them.

From the moment I heard of it, the first thing I thought about was: this could make long range Wi-Fi possible (and therefore meshnets and carrier disruption possible).

Seeing this tech being used by wireless carriers would be cool, but I really want to see it (or something like it) being used in new mile long Wi-Fi standards. Now the Wi-Fi Alliance only needs to make one for us.

Long range Wi-Fi could also make communication between self-driving cars much more feasible (got your attention now, Google?) - although that presents a pretty huge security risk, too, but I imagine the industry already wanted them to be connected to the Internet, so it wouldn't be much worse.

    > Humans use <todo>.

Wireless power transmission will change everything in hardware if someone can pull it off. I'd bet Apple is waiting until they can do low power wirelessly to launch their smart watch.

I doubt that Apple is waiting until wireless power to launch the smart watch, but I agree that it will change everything. And it would make sense for Apple to buy pCell. Of course Steve Perlman worked at Apple before.


Yes. I would much rather Apple spends billions on this than buy silly things like Whatsapp. Article Point out C-RAN being expensive, surely Apple could help or even host their own C-RAN? Since Apple were Originally interested in MNVO anyway.

Don't you still have large power losses from the inverse square though?

Nope (well, yes, as distance -> infinity)

Imagine three boats making waves such that the ripples constructively interfere in some spots and destructively at others.


This will not magically be overcome by constructive interference - each beam will still lose a significant amount of power as function of distance.

Higher frequency RF signals are fairly easy (especially V/UHF) to beamform into a tight area, which can compensate for the loss, by focusing your power in one direction instead of radially. You still lose power over distance as the beam's energy spreads.

I don't believe they will just be able to blob a "networked" 60Hz transmission into a 1cm area - I did not read the patent, but you may be able to do such a thing by varying the phase to induce a 60Hz wave via beat frequencies between transmitting signals, at ONLY the specified area.

Having the necessary +-1Hz resolution on a 2.4GHz carrier (especially one that is not using a hardware connected, phase locked oscillator) may prove very difficult if not impossible. I suspect they may try to use one of the lower scientific bands around 900 or less MHz. HF is also a fairly unregulated option, though tight beamforming may still be difficult. Perhaps a smart application of null steering could create the necessary oscillation - I am still working this out in a simulation.

If this method is correct, keeping everything tight enough to be within US electrical standards will be difficult - there are specific regulations related to in-band power, harmonics, frequency tolerances, etc. Regardless of all that, this technology is pretty incredible and I look forward to investigating it further!

> This will not magically be overcome by constructive interference - each beam will still lose a significant amount of power as function of distance.

Well, yes, and no.

Each beam does "lose" energy compared to the theoretical amount it'd have without interference.

But the power you can transmit into any area inside the cells coverage is nearly constant, and you do not need to put that all extra energy your beans are "losing" into the transmiters, just a tiny part of it (due to errors, delays, object interference, etc).

Why not? Unless I'm missing something, constructive interference is still additive and power loss is still inverse square.

Of course batteries also have power transmission inefficiencies. My intuition though is that wireless power has potential to be more efficient than batteries for two reasons:

- While building out the infrastructue is an enormous proposition, I'd guess that it is much more sustainable to maintain than batteries.

- When the application is propulsion, not having to carry a fuel source (e.g. battery) is huge. On a Tesla Model S, the battery is more than a third of the weight (450kg of 1235kg, according to Wikipedia).

I think with the system they are demoing for communication that is true. They seem to use non directional transmitters of about 5W which cover an area of 'four or five blocks' ( http://www.youtube.com/watch?v=5bO0tjAdOIw&list=PLbOnG9yNO5u... )

A system for transmitting power, which I don't think they have built on any scale, you'd probably want to focus a beam towards the receiver with the power loss being proportional to the amount the beam spreads out.

This is more like radiating narrow transmission beams, not symmetrical spherical transmissions in all directions. There will be losses with distance, sure, but nowhere near as bad as omnidirectional broadcast.

Yes, but constant factors matter.

Consider the simple case of a flashlight bulb by itself, versus in the flashlight (with a reflector). They are both inverse square brightness, but the constant factor in the flashlight case is much smaller.

Similarly this will make coherence areas that will be very small. Wasted power will be approximately inverse-square, but the falloff will be small enough that for short distances you can ignore it.

This is a very nice explanation of not only pCell but of what modern communication systems are like (802.11n has MIMO/diversity but also channel estimation, etc)

People may find this interesting: http://en.wikipedia.org/wiki/Space–time_block_code

Is there any reason this same technology wouldn't work with light? Wouldn't that be a phased light array?

Electronics swithcing speed is the main botleneck to doing this with light. And there are hard barriers that makes this kind of device impossible there.

But different devices with the same results are fair game.

LTE (depending on the band used) can be 10cm down to 3cm long. Visible wavelengths are 300-700 nanometers long. That's about 10,000,000 times smaller!

With seemingly unlimited mobile bandwidth and speed, the cloud gets a new dimension too. Perlman has already proven that even the most demanding applications, realtime games, work well in a streaming scenario even today. With something like this, i could see my desktop in the cloud being accessible from anywhere through a mobile thin client in my pocket. Hook up multiple monitors/keyboard/mouse to it or just use the device screen itself, with computing power only limited by the cloud provider. Technologies like OnLive or PCoIP show thats its possible even today, all thats missing is the infrastructure to support it and pCell could be the answer to that.

While I think charging cell phones and Telsas is a problem, getting cheaper than petrochemical energy is the key. So I have always thought that some sort of space-based energy source was the solution but how do you get it back to earth? 1000's of small solar cell beaming back energy concentrated at one point would be a great solution. How much would this cost would be interesting to find out?

Sort of reminiscent of ESWL, which breaks kidney stones, except in that case they "aim" the sound waves with an elliptical waveguide.

At first I thought the technical challenges inherent in maintaining state for a web of arbitrarily placed transmitters cooperatively targeting thousands of moving devices would make this technology impractical for deployment in the wild, but now I believe. pCell is going to be revolutionary. I can't wait to get rid of my consoles and stream new, triple-A games over it.

USA Federal Communications and other gov agencies EU's for example wanted to move towards telecomms spectrum being a commodity where the device hands off to other carriers seamlessly to handle spectrum scarcity

pCell takes a different tact ..seems that the way to get telcomms to move would be through government mandates tied ot say spectrum sales.

pCell still requires a license to use out in the open. in the same way a fempto cell does.

Sorry, its just beamforming. Its not new, its just marketing.

its not all that different from having an 802.11n MIMO hotspot in your house.

Infact its not all that different from femto cell base stations. You'll still need a license to transmit on that frequency, and you still need a reliable backhaul to dump the data on.

It's different from the MIMO employed in Wifi or "normal" LTE. They do not create a cell per user, neither does a femto cell base station (it only creates a small cell, not a personal cell).

The other, other thing is that inverse square law still applies. along with propagation loss, and the like.

wireless power is literally a waste.

I like to think of it as that conventional cellular is using a broadcast medium for "narrowcast" communications.

If you put it that way it's obvious that we're currently Doing It Wrong, provided there's a practical alternative. It also helps people get over their reflexive "But Shannon says…" reaction.

A friend at CERN commented with this: " This is EXACTLY the same procedure we're using to correct experimental data from distortions introduced by detector effects (unfolding) - http://cds.cern.ch/record/1600778 "

I am pretty amazed by this technology and can't wait to learn more about it. Just a note on the live demo with iPhones — aren't they likely to have buffered the video they need before they were stacked, and therefore it doesn't matter if their antenna connection was interrupted?

"Humans use <todo>" ? Perhaps attention?

Seems we use a variety of mechanisms, not fully understood - http://en.wikipedia.org/wiki/Cocktail_party_effect

Another case of application might be localization.

Isn't this also instant and incorruptable net neutrality?

Maybe it's just me, but that background on the left looks like some monitor that had a seizure.

Applications are open for YC Winter 2020

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