Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
How the electricity markets respond to a nuclear trip (gridstatus.io)
113 points by kmax12 on July 30, 2024 | hide | past | favorite | 84 comments


Something interesting is that with substantial solar and batteries being deployed in the ERCOT market (in scope grid operator for this piece), solar generation in excess of what the grid can consume with load combined with grid forming inverters (vs traditional grid following) can step in when called upon if a traditionally firmer generator (coal or nuclear) trips out. The potential is already there (photons hitting panels), simply turned down to align with grid demand (curtailment), but if the sun is shining and PV is called upon, curtailment can quickly turn into emergency grid support as long as the sun is shining (I am not familiar how long it takes from ISO signal to inverter command).

It would be cool if individual generators shared this curtailment status/reserve in realtime publicly similar to how ERCOT reports real time generation mix data.

Citations:

https://news.ycombinator.com/item?id=40908526

https://news.ycombinator.com/item?id=38848989


> "(I am not familiar how long it takes from ISO signal to inverter command)."

The grid frequency itself functions as a signal: a deviation below 50/60 Hz indicates support is needed, and a deviation above means curtailment is needed (or load added).

Instant frequency-response assets such as batteries typically monitor the frequency independently and respond as required. They don't need to wait for explicit signals from the ISO.


True! But this signal is typically locally monitored and responded to by individual assets ("frequency response" ancillary services). That is distinct from a grid operator, ISO, whomever is orchestrating grid health to call on generating units due to a supply crunch. Tesla Megapacks can respond within milliseconds to frequency and voltage sags (when configured to provide these services), but Autobidder is what orchestrates and pushes out overarching power control strategy to individual assets (for example) [1] [2] [3].

[1] https://www.tesla.com/support/energy/tesla-software/autobidd...

[2] https://www.tesla.com/support/energy/tesla-software

[3] https://electrek.co/2023/09/15/tesla-autobidder-product-330-...


An individual inverter can usually switch in about 4 ms. A Tesla Megapack can go from 0% output to max output in 100 ms. Conversely, gas turbines (the fastest type of traditional power plant) takes about a minute to go from say 40% to 60%. Even in the fastest possible design, there is a mechanical rotor (kinetic energy) that has to change speed.

What I mean to say is that solar and batteries are likely an order of magnitude faster to respond to sudden demand changes. So I would expect a more reliable system when more solar and especially batteries are being added.


> Even in the fastest possible design, there is a mechanical rotor (kinetic energy) that has to change speed.

The point of ramping up the power plant is to make sure that the rotor doesn't change speed (by e.g. burning more fuel to push it harder, because there's suddenly more load on it), and that can happen a lot faster than spinning up the rotor from scratch. Indeed a heavy rotor helps to stabilise the grid "for free" by acting as a flywheel (which, in a way, responds even quicker to the demand change than a battery can).


> An individual inverter can usually switch in about 4 ms.

A quarter of a 60 hz cycle or a fifth of a 50 hz cycle is really fast. For comparison, it takes a current limiting fuse around a half cycle to clear a short circuit current and a GFCI takes around two cycles to clear a ground fault.


I think you’re missing the point how turbine-generator works. The rpm speed has to stay the same, meaning the rotor speed doesn’t change. As long as you don’t have closed circuit you’ll waste that mechanical energy. If you meant starting from stand still position, then you’re right it takes couple minutes to pick up the load.

With that said, turbines responding in couple minutes are more reliable as a baseline when you’re planning load flow of as big as country or wider area. The basic reason is that you have source of energy under your control such as nuclear, water, gas, coal. You cannot have solar, wind as your baseline, I don’t want sound dramatic, but it’s kind of suicidal to do that. Solar’s ramping is not a win when you consider greater scale.


Reliable until it isn’t.

The entire grid is a statistical system where we define the acceptable uptime.

Renewables are as good as any other energy source bringing its own fuel, just need to take the variability into account.

https://www.nytimes.com/2022/11/15/business/nuclear-power-fr...


You’re correct that the system is statistical, and it’s planned accordingly. However, we cannot omit the fact that it’s the running turbine that responds faster to the unpredictable nature of the grid. The backbone of the grid, aka the baseline plants, are extremely responsive to unpredictable nature of the grid at a greater scale, with enough amount of safety margins to bring into service under unusual circumstances. I really don’t see, at least what we have in hand rn, that happening with solar or wind. Without strong baseline you’d experience supply demand imbalance, in engineering terms frequency decay, voltage collapse.


Which means nuclear is the worst companion imaginable. Since nuclear power needs to run at 100% 24/7 to only make an enormous loss.

Dispatchable nuclear power to complement renewables has never made sense.


It's amazing how every comment you make is the most boring ideological propaganda


Stating facts is “ideological propaganda”?

Work with the world, not against it.


Your comments are coming across as lacking nuance ("nuclear is the worst possible...", for example).

Nuclear has a place depending on how you weigh specific factors in your grid design. It's zero carbon. It's hideously expensive, particularly in capex. It's generally quite reliable and its availability is mostly uncorrelated with that of solar and wind. it's modestly dispatchable - you can scale down to 60% or so in many designs. (A little lower but let's be conservative).

If you place high weight on zero carbon, nuclear is an (expensive) way to get through the night. It can work pretty well in a grid mix if your grid is large enough that the loss of one nuclear plant isn't a really big chunk of your power supply (since, obviously, you want enough redundancy to handle a certain fraction of generation failures at peak load).

Are solar+wind+batteries on a much better trajectory? Yes. But batteries are not there _yet_ for 24x7, though I think we all hope they will be in the reasonably near future.


He does not care about these arguments, what matters is that there is no room for nuclear power.

In a past discussion I talked to him about how one of the important things to do was to diversify, as China has a lot of influence on the whole renewable sector (solar, batteries, etc.)

Needless to say, that's not a problem for him. For him to hope that batteries are the future is already a sure thing, without the slightest doubt.


Hideously expensive and any plant announced today will not be online in time to have any material effect on our fight against climate change.

Which means funding diverted from renewables to nuclear will prolong our fight against climate change.


> Hideously expensive and any plant announced today will not be online in time to have any material effect on our fight against climate change

False, we have 26 years to decarbonize, all the time it takes to build any number of nuclear power plants in any country in the world.

> Which means funding diverted from renewables to nuclear will prolong our fight against climate change.

We can say the same thing about renewables. Then come and tell me you are not ideological... Where is the mathematical certainty that batteries at scale will be available everywhere and for everyone by 2050? If you come from the future, prove it to me and I will agree with you.


About 20 years from being announced. Compare with renewables taking 1-5 years depending on if it is solar or offshore wind.

Say 5 years for renewables.

This means that investing in nuclear will have 15 years of cumulative emissions before anything is curbed.

Meaning, even if the renewable options ends up solving only 80% of the problem it will take until somewhere 2080-90 for the “perfect” nuclear solution to have less cumulative emissions.

Even if renewables are completely unable to solve the entire problem we can invest in them and then in 2060 and still be ahead of nuclear power, and then choose it as the final solution.

Today it is simply lunacy proposed by the fossil fuel industry or people looking for the perfect solution rather than piecemeal solving the issue.


These are lies you tell yourself and how you want people to see you. In a past discussion, you concluded by saying that those who support nuclear energy also support fossil fuels.

These are your ideological premises; you don't care about creating a better world, nor are you interested in facts and problems. You only care about your vision of things and making it prevail over others.

There's no need to know anything else to make any of your comments irrelevant. It's no coincidence that you are in every nuclear discussion, asserting how much you are against it.


Given how the rightwing conservative politics have shifted from pure climate change denial to harping nuclear as the non-solution to prolong our reliance on fossil fuels the link is clear.


Nuclear power derives its energy from the binding energy of ultra-heavy atomic nuclei, not fossil fuels. This energy can then be used to power electric cars, avoiding fossil fuel consumption. Furthermore, unlike wind and solar energy, nuclear power generation is not tied to the vagaries of weather, meaning that it doesn’t require burning millions of cubic meters of natural gas every overcast, calm day. For this reason, Russia historically was the largest single benefactor of Germany’s green euphoria: it prolongs fossil fuel reliance.

So often, emotional thinking leads to conclusions that are opposed to reality. You really have to watch out for it, if you want good results.


IANAEE (not an electrical engineer), but doesn't a gas turbine spin at a constant speed to drive a 60Hz generator, regardless of whether it's running at 20%, 60%, or 100% of capacity?


When there’s a load imbalance on the grid (more load than capacity), the turbines physically slow down as inertial energy is extracted from them. This causes the grid frequency to drop. It takes some time to ramp up production and speed up the turbine etc.


The maximum slow-down of the turbines (before the generator trips off-line, removing the load on it) is far less than you seem to assume. The article's graph shows the 60.00Hz grid frequency dropping...all the way to 59.92Hz. That's 0.1333%.

For an on-line gas turbine, the time to ramp up production is the second or few needed for the automated controls to open the throttle on the "Gas IN" pipe. It's basically a natural gas-burning turboprop jet engine, with the propeller replaced by a generator. (Yes, this can be less efficient in a combined cycle plant.)


The more interesting thing about rotating power plants is that they are routinely destroyed by transmission outages, because when a rotating generator is suddenly disconnected from its load, there are infinity terms in the equations that govern its motion and infinity isn't a thing you can resist. For steam turbines the control system has to slam the valve shut on the steam, otherwise the machine would overspeed, and closing that valve destroys some sacrificial part of the steam plumbing (hopefully). Steam power plants have to be inspected and repaired after disconnects and this is one of the numerous reasons why fission kinda sucks on the reliability front.


While the infinity terms may be a useful way to think about how the system loses stability, I think it's better to think of it in terms of energy: at any given time there's some amount of energy that's been injected into the system in the form of hot gas which hasn't been converted yet. Even a few seconds worth is a lot of energy, and if the electrical load is removed, that energy has to go somewhere.


I don't see why there can't be load resistors (like locomotives use) to gracefully handle these sorts of situations.


An SD40 locomotive can dissipate about 500kW that way, so you would need the equivalent of 2000 locomotives worth of resistors and fans to sink the 1 or 2GW or more that a fission power station produces.


If you wanted to do so continuously. If you're just handling an impulse you don't need so much, but it's gotta absorb a lot of energy quickly.


Well if you can solve it, I imagine there's a lot of money to be made - and yet it hasn't been solved.


This is true only if the protection does not operate as designed. They are NOT routinely destroyed by transmission outages. You have been misinformed. Rotating machines cannot overspeed instantly, as they have inertia.


Dozens of power plants blew up steam rupture disks during the 2003 northeast outage. EPA, FERC, and NYISO all mention this in their reports.


How many Tesla Megapacks are needed to replace a nuclear plant temporary outage? And for for long can they provide that power?


Most curtailment IIRC is due to insufficient transmission capacity. I doubt curtailed solar can be called upon in an emergency unless the emergency is located very close to the curtailed solar.


It's a fair point (why curtailment is in effect), and I think speaks to the fact that more granular and timely data is needed wrt all nodes and transmission segment within the system. Also a call for more batteries everywhere between generation and load.

With regards to transmission congestion, that is easily fixed with installing batteries at currently storageless renewable generation facilities (the batteries then charge with excess solar, and can continue to discharge after the sun sets or the wind dies down, maximizing transmission utilization temporally). The Inflation Reduction Act also enables those batteries to charge from utility side if needed, whereas before they could only charge from the renewable generation (AC vs DC coupling).


Batteries are extremely expensive per megawatt, not very durable, require carefully controlled temperatures, and their manufacturer and recycling extract a tremendous cost from the environment. For non-mobile usage, batteries shouldn't be seen as any kind of viable solution at scale.

However, there are other ways to store energy; unfortunately, most involve converting electricity to another form of energy such as potential (gravitational) energy, like pumping water uphill or lifting heavy weights. These also have relatively little long-term environmental cost. Unfortunately, they're a bit more inefficient (but so are batteries, relative to some other forms of stored energy such as fossil fuels).

It'd be interesting if we could find some ways to convert landfills or other urban blight issues into a durable energy store without poisoning the environment.


> For non-mobile usage, batteries shouldn't be seen as any kind of viable solution at scale.

Lithium-ion, sure, but aren't there a whole host of other battery chemistries that are basically too big / too heavy to put on vehicles but a lot cheaper so well suited for stationary storage?

Are they all still at the research phase and so currently more expensive than the decades-of-learning-curve lithium-ion?


Lithium-ion is the cheapest form of stationary storage for the sub-8 hour duration niche. The vast majority of battery storage being deployed is lithium-ion.

Sodium-ion is the second largest contender, with a few pilot facilities opening in China recently, but it will be a few years before it eclipses lithium-ion.


The vanadium redox flow battery. https://en.wikipedia.org/wiki/Vanadium_redox_battery

> For several reasons, including their relative bulkiness, vanadium batteries are typically used for grid energy storage, i.e., attached to power plants/electrical grids.

    VRFBs' main advantages over other types of battery:

    no limit on energy capacity
    can remain discharged indefinitely without damage
    ...
    wide operating temperature range including passive cooling
    long charge/discharge cycle lives: 15,000-20,000 cycles and 10–20 years.
    low levelized cost: (a few tens of cents), approaching the 2016 $0.05 target stated by the United States Department of Energy and the European Commission Strategic Energy Technology Plan €0.05 target
(and yes, there are disadvantages too)

See also NPR's story: The U.S. made a breakthrough battery discovery — then gave the technology to China https://www.npr.org/2022/08/03/1114964240/new-battery-techno...


How is 10 - 20 years "sustainable"?


We’re currently experiencing a Cambrian explosion in battery tech. As the technology matures, and we establish a closed loop ecosystem to build and then recycle these systems, longevity can improve over time. To get better at something, you must first suck at it, and 10-20 years is not an immaterial service life for an asset that just sits and hums with no moving parts.


I agree if we were talking about a motor or a pump, but it seems like batteries basically devastate the environment every time we make one, and doing that millions of times every ten years is probably not great. (But I don't know anything about that specific battery technology.. perhaps it's just saltwater and two dissimilar metals.)



Not a single thing you said about batteries is true. Pumped hydro is in no way competitive with batteries for most locations. In the future it is likely they won't be competitive in any location.


Better still; the amount of deployed batteries world wide is projected to overtake the amount of deployed hydro this year. Pumped hydro is barely growing. Battery capacity is growing exponentially to eclipse it this year. That's driven by pure economics. Cheaper, better, faster, etc.


Neither of these points are conflicting with my argument. Just because batteries are being deployed it (and I agree that it's cheaper and faster, but not necessarily better on all axes) doesn't mean that their manufacturer is not damaging the environment.

In fact, battery manufacture is not damaging the environment in most places where first-world people live, so perhaps they just don't care, but I think that's pretty sad.


If that wasn't clear, I wasn't trying to challenge your argument; just adding to it. And you are right by adding more arguments to the pile. Not disagreeing at all.


> Batteries are extremely expensive per megawatt, not very durable, require carefully controlled temperatures, and their manufacturer and recycling extract a tremendous cost from the environment. For non-mobile usage, batteries shouldn't be seen as any kind of viable solution at scale.

None of this is accurate. I encourage you to update your mental model with recent data. Citations below for your convenience. AMA, global energy transition is my passion.

https://www.utilitydive.com/news/batteries-texas-consumers-6... ("Utility Dive: Batteries saved Texas consumers $683M during 2-day January freeze: Aurora Energy Research")

https://news.ycombinator.com/item?id=40919052 ("HN: China's Batteries Are Now Cheap Enough to Power Huge Shifts")

https://news.ycombinator.com/item?id=40601878 ("Lazard: IRA brings LCOS of 100MW, 4hr standalone BESS down as low as US$124/MWh"

https://news.ycombinator.com/item?id=35513612 ("HN: The biggest EV battery recycling plant in the US is open for business")

https://www.bloomberg.com/news/articles/2024-04-24/battery-r... | https://archive.today/OjA91 ("Bloomberg: Battery Recycling Shatters the Myth of Electric Vehicle Waste")

https://www.lazard.com/research-insights/2023-levelized-cost... ("Lazard: 2023 Levelized Cost Of Energy+")

https://raokonidena.substack.com/p/history-of-10000-cycles-o... ("History of 10,000 cycles or 10 year warranty for Battery Energy Storage System (BESS)")

https://www.nrel.gov/docs/fy23osti/85332.pdf ("NREL data shows BESS asset life is 15-20 years.")

https://web.archive.org/web/20240728011101/https://www.eia.g... (US EIA; gray installations are new battery storage planned for deployment over the next 12 months, June 2024 through May 2025)

https://www.eia.gov/todayinenergy/detail.php?id=61202 ("US EIA: U.S. battery storage capacity expected to nearly double in 2024")

https://www.eia.gov/todayinenergy/detail.php?id=61424 ("US EIA: Solar and battery storage to make up 81% of new U.S. electric-generating capacity in 2024")

https://www.woodmac.com/blogs/energy-pulse/battery-storage-b... ("Wood Mackenzie: Battery storage begins to play a key role for US grids")

https://www.nytimes.com/interactive/2024/05/07/climate/batte... ("NY Times: Giant Batteries Are Transforming the Way the U.S. Uses Electricity")


> global energy transition is my passion

Seeing what you write, it seems more like an obsession disguised to support your ideologies


A recent article asks:

"Why did the U.S. miss the battery revolution?"

https://www.noahpinion.blog/p/why-did-the-us-miss-the-batter...

About half the comments on the article were Americans saying very similar things to your comment and denying there was a revolution to have missed, which kind of answers the question posed.


They're the cheapest form of storage after recent price declines, even cheaper than pumped hydro and compressed air.


The secure authentication of all those nodes concerns me. Are these old scada systems coommunicating over plaintext rs232 or similar? Is it something running crowdstrike?


The transmission lines run to solar farms should be able to take 100% of the output of the farm and then some otherwise the farm was over built and wasted money.


At short distances from the solar farm, yes. But it's not 100% in every direction for arbitrarily long distance. At some point, you assume the energy will be tend to be used sort of near where it's generated.

To put it another way, if you build solar between city A and city B, would you build it so it can still be fully utilized even if city A stops using any power and city B wants all of it? No, you assume city A is always going to need some power.


Generally they connect directly to large back bone transmission lines that carry power far beyond the local area though not directly to more balkanized power zones. On a large scale yes if critical junctures go out the rest can't take the full load but that's different than a single plant being overspecced for the transmission capability it's connected to.


That's not necessarily optimal. For home installs, you can overbuild panels because they're cheap compared to the inverter. Then you curtail sometimes at midday and have extra energy on cloudy days and in the morning/afternoon. Turns out that's more cost effective than sizing the panels perfectly. The same logic could apply to utility farms, because transmission lines can be expensive. I don't really know myself since I don't work in the industry, but I would not be surprised if they slightly overbuild vs. the transmission line capacity.


It definitely applies to utility builds. It applies even more if the DC power from the modules can also be used to directly charge batteries.


If by output you mean the maximum output of the inverters. The maximum output of the PV modules can be higher, particularly if the field has integrated batteries.


By this definition all reserve capacity is wasted money. That is clearly false, as demonstrated by the event in this article _not_ leading to failure and harm.


I'm not talking about reserve capacity at the grid level. I'm talking about excess capacity at the individual generation plant level that exceeds the grids capacity to take in. If you can't output the energy onto the grid the only benefit is for local maintenance and you don't need huge amounts of excess capacity to solve that and that excess doesn't help in the event of a large base producer like a nuclear plant going offline because it can't get onto the grid!


I agree with your use of the word "should" here. But that unfortunately is not what happens in reality.


Grid following inverters can do something similar, although their response time is usually limited to the grid operator's automatic generation control (AGC) communication cycle, which is typically on the order of 4 to 6 seconds. And this requires plant controllers, and markets/contracts, to be setup for it.

See https://www.nrel.gov/docs/fy24osti/86932.pdf. It deals with estimating reserves from these types of resources, but also talks a bit about general considerations, and references other good papers and demonstrations.


Something less interesting and more annoying is that in power generation surpluses instead of doing this, the government can just mandate that you buy power from them instead of get it free from the sun.

Because a few hundred thousand chips in solar systems costs less than an entire network rethink.


I helped write this! The data behind all the graphs in this post is here: https://www.gridstatus.io/dashboards/ercot-nuke-trip-july-20...

If you like grid data, we have a lot more info for the entire country here: https://www.gridstatus.io/live


Those links are very informative. Can you elaborate on Ancillary Services Monitor and Energy Storage Resource Action dashboards? What’s the installed capacity that supported?

Also your original post link states event happened at 7:02 AM, your links here points to 8:05 AM. Can you explain this?


You may need to check the time zone as some pages automatically adjust to local time of your browser. It happened at 7am CT.


That makes sense, I'm in EST.

Just to clarify and understand what happened, I believe right after the trip some generators all around the grid picked up the load (unless UFLS was activated) immediately (around 7:03), we can call these generators support system. Then around 7:05 batteries kicked in with 468 MW, as a support to support system.


Where does the 1500MW of heat the reactor is producing go, immediately after the generators are isolated from the grid?


Enthalpy of vaporization for water is very very large.

The amount of heat that can be dissipated by evaporating water is kind of incredible.


Thanks, that's what I was asking.

(It's obviously not a meltdown situation!)


It seems like you might really be asking about a meltdown scenario.

The reaction can be slowed with control rods, which stops/minimizes the heat from being generated. The previously generated heat still needs to be handled, however (by evaporation).

https://en.wikipedia.org/wiki/Nuclear_reactor

A meltdown occurs when the reaction can't be slowed down through normal means because the safety systems fail.

For example, Fukushima and Chernobyl:

https://en.wikipedia.org/wiki/Fukushima_nuclear_accident

https://en.wikipedia.org/wiki/Chernobyl_disaster

However, those safety systems are obviously designed expressly to prevent such a disaster. For example, control rod systems are often designed to be fall (via gravity) into the reactor in the event power fails. These failsafe systems are typically very reliable, in the absence of other external events (such as the flooding and earthquake that took place in Fukashima.)


The control rods don't stop decay heat from being produced. That's what happened with Fukushima.

Also, nuclear poisons (neutron absorbers) build up after the reactor is shut down. After the control rods are withdrawn it takes a few days for the poisons to be burned up and the reactor can resume power production. I think they call that poisoning out the reactor.

https://en.wikipedia.org/wiki/Neutron_poison


When the Northeast power grid went down down in 2003, the Bruce Nuclear Power Development was almost entirely kicked off the grid. It was able to keep ticking over at a few percent of power output though because of a feeder line that went North. This prevented the reactors from poisoning out, and allowed them to come back faster than if they had poisoned out.


> These failsafe systems are typically very reliable

From the RISKS archives:

"A fail-safe system fails by failing to fail safe."


Instead of that energy being directed to the turbines to convert the thermal energy into rotational energy (and thus it to electricity), the reactor begins cooling from the primary reactions (but not completely since secondary reactions/decay still are going) since there is no reaction heat (reactor idled). All of that excess residual heat contained in the mass of the fuel and reactor water is carried away by the cooling water loop, and either out to the cooling tower where it evaporates and releases its energy, or through the next stages of the loop into the cooling supply, most likely a neighboring river or lake.


In general, any sort of "boil water to steam, use steam to drive turbines" system can vent the hot steam from the boilers to the atmosphere. There's a big reserve tank of cold boiler feed water, to replace the water you're no longer getting back from the steam condensers (attached to the "OUT" steam pipe on the turbine). So the boilers will keep soaking up just as much heat, while the engineers scramble to reduce the heat coming in from the burning wood, or burning coal, or burning oil, or fissioning atoms, or whatever.


The reactor is turned down and rapidly stops producing that much heat. On site backups keep the pumps running to continue cooling the residual heat. There's a long slow tail as lower energy fusion chains tail off towards lead but the vast majority of the power comes down quickly.


I would assume the cooling loop is still running on generator and/or backup utility feed power, so the heat from the reactor would go into the water of the cooling loop, and then the heat would be rejected into the atmosphere when the cooling loop water evaporates in the cooling towers, with the leftover heat being discharged into the river/lake/cooling water source.


Water? Same place it was going


I mean, at least some percentage was going down the wire. That’s why we do it, right?


It would be funny if we were doing the whole song and dance just to amuse the electrons.

If the turbines are 50% efficient, then the heat load on the cooling towers would double, which I have to assume is WELL within design limits.

You have made me wonder about turbine overspeed and steam bypass.


I think the nodal pricing structure in the Texas grid is extremely conductive to an efficient electricity market.

It incentivises the grid-stabilizing deployment of batteries.

Texas is really the market to watch and learn from.


I know Texas wants their independence & politics etc but its a bit silly not to have at least an emergency linkage to the bigger grids around to help catch falls like this. Even a fairly modest 500MW or whatever.


They have interconnects with Mexico. They relied on a 300MW interconnect during the 2011 winter storm outage.


The Texas powergrid is so damn cool.




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

Search: