Alphabet spins out Taara – Internet over lasers
201 comments
·March 17, 2025jauntywundrkind
krisoft
> Line of sight free space optics can be immune to many many forms of jamming.
I’m a bit of two minds about this. Obviously jamming resistant high bandwidth communication enables some scarry possibilities.
But the lack of it is what drives and will drive militaries around the world to put more and more autonomy into weapons. It doesn’t matter what kind of treaties we write on paper to prohibit technologies. During a war if your drones/loitering munition are less effective than those of your enemies because your control signals are jammed you will give in and make your weapons find their target without that control signal. That leads to an arm race of ever more sophisticated autonomous weapons. That is scarry for many reasons, and probably a worse outcome for all of us.
On the other hand if communication is possible that puts a leash on this dynamic and ensurers that a human mind can remain in the loop. So… maybe being better at jamming resistant communication is actually better for humankind?
femto
> On the other hand if communication is possible that puts a leash on this dynamic
I'm a bit more pessimistic than that. I think the driver for autonomy will be that the speed at which things happen on the battlefield. People being attacked with automated weapons might not be able to make response related decisions fast enough. The automation will be in place to enable a rapid response. It will become an arms race involving speed of attack and response. It will be the military equivalent of high-frequency trading, involving things like swarms and directed energy weapons.
ethbr1
The phrase you're looking for is OODA loop, popularized in the 80s.
https://en.m.wikipedia.org/wiki/John_Boyd_(military_strategi...
BurningFrog
I got seriously terrified by reading a PKD SF story at 15 about the few surviving humans hiding from war drones still hunting people long after the war had ended.
stevenwoo
Second Variety, https://en.wikipedia.org/wiki/Second_Variety I've commented this before but it is closer to Terminator than the Harlan Ellison story IMO.
falcor84
I haven't read that one, but it reminds me of Black Mirror's "Metalhead", which absolutely terrified me.
umvi
> Line of sight free space optics can be immune to many many forms of jamming
The most powerful weapons on earth already are immune to jamming. ICBMs use celestial navigation (pictures of the stars) to course-correct, which is a form of navigation you cannot jam.
aerostable_slug
To risk being pedantic, US SLBMs (Tridents) indeed use stellar updates for their inertial guidance system.
Our ICBMs have no update mechanism at all. Once they're out of the silo, they're completely on their own.
HPsquared
ICBMs would be more confident in their starting position and orientation. SLBMs would necessarily have more uncertainty in the initial conditions.
BurningFrog
The BM part means "Ballistic Missile", which sounds like they're basically "thrown" into a parabolic path towards their target, like a falling rock.
Or am I over-interpreting the name?
beardedmoose
The missile knows where it is at all times. It knows this because it knows where it isn't. By subtracting where it is from where it isn't, or where it isn't from where it is (whichever is greater), it obtains a difference, or deviation. The guidance subsystem uses deviations to generate corrective commands to drive the missile from a position where it is to a position where it isn't, and arriving at a position where it wasn't, it now is. Consequently, the position where it is, is now the position that it wasn't, and it follows that the position that it was, is now the position that it isn't.
In the event that the position that it is in is not the position that it wasn't, the system has acquired a variation, the variation being the difference between where the missile is, and where it wasn't. If variation is considered to be a significant factor, it too may be corrected by the GEA. However, the missile must also know where it was.
The missile guidance computer scenario works as follows. Because a variation has modified some of the information the missile has obtained, it is not sure just where it is. However, it is sure where it isn't, within reason, and it knows where it was. It now subtracts where it should be from where it wasn't, or vice-versa, and by differentiating this from the algebraic sum of where it shouldn't be, and where it was, it is able to obtain the deviation and its variation, which is called error.
tryauuum
that is so romantic. Rockets looking at the stars one last time before obliterating $cityname
HPsquared
Observing distant nuclear fusion, before unleashing it locally.
freddie_mercury
This makes me wonder how many fake stars an adversary would have to put up (and for how long?) in order to confuse a celestial navigation system.
echoangle
This form of navigation is probably only accurate enough for nuclear weapons, you’re not going to get meter-Range CEPs with that. You probably have to select a city you want to hit.
aerostable_slug
This is a very good read on the state of the art when it comes to submarine-launched missile accuracy, which are presently inertially guided with a stellar update during flight.
https://thebulletin.org/2017/03/how-us-nuclear-force-moderni...
cogman10
While you probably can't target a cigarette butt on the street, you could definitely hit a building. Especially if the ICBM is paired with image recognition (which it has already for star nav) and/or backup positioning mechanisms like cell tower locations or well-known broadcasting tower locations (think television stations).
twobitshifter
well there is this now as well - you don’t need to use the stars when you already have detailed aerial images of the earth - https://www.spectacularai.com/gps-free
dzhiurgis
Quantum navigation is likely to be more precise.
bsimpson
Can you imagine if someone just turned off the sky?
runako
Spin by Robert Charles Wilson is based on this premise.
faraaz98
Who are you? Cixin Liu?
charliebwrites
Damn guess we have to blow up the stars
…you know, for defense
echoangle
Or create a lot of artificial ones to confuse the missile.
LoganDark
Aperture Science would do it.
spwa4
I think you're too late. Pretty sure the military has line-of-sight free-space-optics to satellites since at least the second Iraq war.
One can only wonder how small the "receivers" (routers, really) have gotten by now.
rhubarbtree
Cubesats carry laser comms if that's what you mean.
echoangle
Those work in vacuum though so you need a bit less laser power. The atmosphere attenuates your signal if you’re doing ground to ground or ground to aircraft links so you probably need a bit more laser power. But I agree that that’s probably not the difficult thing about the whole system.
cogman10
LOS communications has been around for a long time now in the form of microwave towers. [1]
But further, jamming is still doable, just not with a portable electronic device. Stir up some dust and all the sudden coms are down.
miohtama
I was doing communications in military. For temporary networks, we used microwave links and they requires line of sight. The reason to use direct link is that missiles can be easily targeted to any radio source you can hear and take down the network.
cogman10
Yeah, I was aware of this because of work with telecommunications in rural locations. Nobody is burying a cable over a mountain, that's too expensive and too much work. So instead, they'll setup microwave towers. Being on a mountain has the benefit that you are already in elevated locations that are easy to shoot a phone signal across the range.
cameldrv
This is something the U.S. military has been working on for some time. I know of at least one project that's very advanced: https://www.ga-asi.com/multi-mission-payloads/lac12-pod
nostromo
This is already in use with the US military.
lutherqueen
How is better communication scary for military? The weapons themselves are
FlyingBears
Jam resistant comms are critical for drones, and other precision weapons and their infrastructure. Even if line of sight is interrupted modern drones can return to signal nowadays, relay information, and return to target with corrections. You may not need optical cable anymore
HeatrayEnjoyer
I expect drones will become fire-and-forget in two to three years. They won't be jammable because the pilot is in the drone.
ecshafer
Should the military not have internet? If laser based internet is better than satellite or microwave or wireless (I assume the military uses these three). Then isn't that good?
geek_at
[flagged]
r14c
That's the interesting paradox of technological development. If it's out of the bag, then it's out of bag, but that cuts both ways. You can never control who will gain access to advancements. Even if you trust the custodians now, you don't know who they will be in the future, &c. Eventually you end up at a fork: either pursue and endless technological arms race or find some way to build lasting cooperative and peaceful relationships. As they say, the surest way to destroy your enemy is to make them your friend. Still a massive challenge, but presumably that's why empires eventually always fail. They don't know how to make friends.
EvanAnderson
RONJA[0] with lasers? >smile<
Free space optics always seemed like a neat idea. For space-based communication, particularly if your "mission" involves as little stray emission as possible, I would think free space optics would be a win.
I would assume there's more error correction, but otherwise I wonder how dramatically this differs from modulating light on a fiber. It seems like a similar problem.
proxysna
Exactly my first thought. Had this link in my bookmarks for an eternity. http://images.twibright.com/tns/1208.html
fsh
This section from the marketing blurb doesn't sound too promising:
When atmospheric conditions disrupt the light, our adaptive rate and hybrid architecture maintains the connection, with minimal downtime.
In the long run, all these wireless technologies (satellite or optical/microwave terrestrial links) will have a very hard time competing with simply laying down some optical fiber.
gymbeaux
The economic burden usually falls on governments, so, like StarLink, Alphabet is probably hoping for some of that sweet, sweet government subsidy/grants for military applications.
mattlondon
Some of their use-case they are crowing about on their site cover temporary things: back haul for major-but-temporary events, tethered-drone-mounted units for emergency disaster recover where a cell site is taken out etc. Those are the sorts of things where laying fibre 20km for use for just a day or two just isn't going to happen, but a temporary laser link that you can get up and running in a hour or two would be great.
lelandbatey
What kind of data rates and distances are they talking about that isn't served by existing products? For example, you can buy a 20km range, 2Gbps wireless point to point link for a flat $3000 today: https://store.ui.com/us/en/category/wireless-airfiber-ptp/pr...
What they mention in the article is up to 20Gbps, but they'd have to be pretty dang cheap to out compete just buying 10 of the existing options.
cycomanic
The issue is that you can't put 10 of your 2 Gbps wireless links next to each other. You quite possibly end up with < 2 Gbps as interference kills your signals (unless you put the transceivers so far apart from one another that you sort of defeat the purpose). That said there are other wireless solutions that can get you > 10 Gbps over > 20 km already (not sure about 20 Gbps, but I wouldn't be surprised). The issue is available spectrum, i.e. you can't just setup the link, because the spectrum doesn't belong to you. Not a problem for optics.
gmoot
Elsewhere in the thread it suggests ~$30k for one link. Which is exactly in line with buying 10 of the ubiquiti devices.
But I think you would need 20 of them, 10 on each end? Plus extra install, networking equipment, etc. Which would make Taara significantly better.
delulu42
[dead]
culi
That's not the market they're going for though. They're more of a competitor to Starlink
There's also obvious applications to places where weather is more predictable. There's plenty of areas and small towns in the Great Basin region that have basically no internet. This would be a quick and easy way to set those places up with internet with more reliability than something like starlink
pdabbadabba
But why would these not be places already served by terrestrial wireless internet service providers? It seems like it would be much easier and generally more attractive to serve locations like this using, for example, 5 GHz.
Normally, the lack of (near) line-of-sight is one of the biggest limiting to those sorts of deployments, but that would also have to be solved for any place being served with FSO.
fsh
The problem with selling inferior technologies is that sooner or later people are going to stop using them (even in the Grad Basin region). Not exactly a recipe for success.
walrus01
a point to point terrestrial bridge large piece of equipment that costs $5,000+, needs a professional to install it, and works on either free space optics or V-band or E-band radio is not in any way a competitor to starlink. It's more a place to take a 1 to 10 Gbps ethernet connection as a link between two towers or roofs that can 'see' each other as an alternative to where laying fiber may be cost prohibitive or would take too long to build (or both).
Assuming this thing doesn't utterly fail in rain at a moderate distance, this would be something you use to feed a POP which then redistributes service to end users by some totally other technology (5/6 GHz band PTMP radio system, GPON, XGSPON, G.fast on copper, docsis3/docsis3.1, etc)
slowtrek
"... very hard time competing with simply laying down some optical fiber."
You end up learning this in your own home. Some things are fine with a wired ethernet connection, it's really only my laptop and phone that use wifi.
adrianmonk
You can say the same thing about running wired ethernet to your TV in the living room. It's simpler and more reliable than wifi. But wifi is much easier and quicker to install. Which one do most people use?
fsh
For most users (me included), there is zero difference in user experience between using wifi or Ethernet for their TV. Otherwise, running wired Ethernet would probably be a lot more popular.
TeMPOraL
So you think. You may be right, but most users won't even realize that a good chunk of their "buffering" / "Internet is slow today" / "Netflix is broken today" problems might just be a WiFi issue, and it would go away if they used a wired connection.
darksaints
Optical fiber is absolutely the simplest and best option for almost any form of long distance connectivity. Maybe this technology will become cost/performance competitive in about 15 years after the HFT firms have invested billions trying to extract an extra cent out of our financial markets.
idkwhatever
[dead]
refulgentis
It's interesting at 36 to look back at what I think would disrupt connectivity a decade ago:
- Google Fiber (it wasn't possible to do it cheaper than incumbents, so it devolved to standard incumbent x why would 40% margin company invest billions to get Comcast's peak profit margin of ~15% profit)
- Starry Internet (too expensive to build out, I have it and it's good, but the company certainly didn't scale)
- 5G in general (strictly inferior to incumbent, speed isn't faster, latency is higher, not as reliable)
It's hard for me to wrap my mind around why this would work at all, sounds like a more-susceptible-to-bad-conditions version of Starry.
I keep wondering how people make Starlink work, my understanding is the connection degrades then stops then reconnects every...idk, 5 minutes? as the satellites go overhead.
Workaccount2
The key breakthrough for 5G was allowing ~10x the number of devices to connect to a node compared to 4G. 5G is what allowed the toppling of data caps that was by far the #1 consumer complaint for years. 4G just couldn't handle heavy loads well, so data caps were needed to constrain demand.
Teleco's aren't going to say this out load, but it's the real reason why they were so celebratory about 5G, despite it coming off like just a renamed 4G to the average user.
jampekka
> 4G just couldn't handle heavy loads well, so data caps were needed to constrain demand.
In many parts of the world uncapped data has been the norm since around GPRS.
stnmtn
Why would they not be loud about it? I think "We've built out 5G so we can get rid of your data caps!" is a message any telecom would want to broadcast out, unless I'm missing something
calebio
> I keep wondering how people make Starlink work, my understanding is the connection degrades then stops then reconnects every...idk, 5 minutes? as the satellites go overhead.
That is not a correct understanding for how the Starlink network behaves today[0]. While I can't speak for using it outside of the U.S., I have not faced any interruptions outside of a few times during very severe weather.
[0] in the early days of the constellation, there were sub-second or a few second drops when there was no satellite overhead. But this dropped off very quickly once the constellation size increased.
refulgentis
I see, tyty (been wondering for quite some time)
johncalvinyoung
As I deployed Starlink in an extremely obstructed spot last year for a few weeks, where multi-second dropouts were quite common... it impressed me JUST HOW MANY satellites they have up there, and just how usable my dish was despite only having ~60% of its field of view clear. It's switching satellites much more often than every five minutes.
The built-in obstruction mapping tool quickly demonstrated that though each satellite represents a tiny slice of sky... over the course of the day you're seeing a vast number of satellites at a high variety of spatial angles and orbits.
I wouldn't recommend that obstructed situation to anyone (and it's going in a much clearer location this coming summer) but the users I was supporting reported it a far far better solution than the 4G LTE they'd been depending on prior. Not a patch on fiber, but a great solution for an awkwardly remote property.
cbg0
> 5G in general (strictly inferior to incumbent, speed isn't faster, latency is higher, not as reliable)
I'm guessing this is a US thing? In Europe, 5G is definitely faster while latency is on par with 4G. YMMMV between EU countries though.
refulgentis
You're right, it's definitely better than 4G, my wording was unclear, more in the sense of "Would I make this my home ISP?" than "how did 5G go?" (I would have thought cell providers would have 20-30% of the market now, ah, the follys of youth...)
miloignis
From my perspective, Google Fiber 100% disrupted connectivity - it woke the incumbents up and made them offer competitive Fiber. In that sense, they succeeded! My last three connections from my last three ISPs have all been gigabit (one of which was Google Fiber, easily the best internet I've ever had). I think they're expanding again, too, though I wish they had stayed as aggressive with rollout as they started.
jerlam
Ironically, Google Fiber purchased a wireless provider - Webpass - back in 2016 which is deployed in parallel to their fiber offerings.
refulgentis
That's a really good point, back home, Verizon didn't bother with Fios investment until then.
woah
It's crazy that almost every house is able to be attached to a pipe carrying high pressure water that will flood if it is broken or attached wrong, thick wire carrying high current that will shock you, a pipe containing explosive gas, and a six inch cast iron pipe full of poop, but adding one more connection to a tiny thin strand of glass wrapped in plastic is too expensive.
srameshc
I thought this was a link to X(twitter) but this is a Google X
black_puppydog
Big tech deserves to burn for their evolution in naming alone.
bix6
Bmw would like a word: BMW M760Li xDrive V12 Final Edition
rcpt
Is that a gaming monitor?
mdaniel
And also squatting on the i3 name because one of these is not like the other https://duckduckgo.com/?q=bmw+i3+2020&iax=images&ia=images vs https://duckduckgo.com/?q=bmw+i3+edrive+35l&iax=images&ia=im...
I mean, seriously, there are more numbers available
elromulous
Fwiw, Google X predates the Twitter -> X rename by more than a decade.
boguscoder
Ironically Xerox Parc experimented with network over laser more than 5 decades ago
light_triad
Apparently the biggest problems are line of sight interruptions and cost:
> The team has figured out how to compensate for potential line-of-sight interruptions like bird flights, rain, and wind. (Fog is the biggest impediment)
> “It’s fast and reliable but quite expensive.” He says he spent around $30,000 for the last light bridge setup he bought from Alphabet for testing.
Interesting that Meta was working on similar tech but abandoned the project:
Google’s Taara Hopes to Usher in a New Era of Internet Powered by Light
https://www.wired.com/story/plaintext-google-taara-chip-inte...
More on Meta's internet via lasers project:
https://www.wired.com/2016/01/facebook-zuckerberg-internet-o...
pmcf
Line of sight issues are simply a wattage issue. A gigawatt laser is impervious to rain, a bird, a flock of birds, a bird and the tree it sitting in. Probably the entire forest. Let’s just say there are some solutions well in hand.
rpmisms
I like your first sentence as a truism.
kfrzcode
It does have a lot of potential.
"At once, the planets stood between us, forever occluding what we yearn to see.
For existence itself could not repent, the time we spent, and yet
Only whence wielding wattage whole had we defined;
-- an oculus toward which we'd partake of thine.
To space and time, no morphisms apply, no longer shall ye escape our Eye."
walrus01
$30,000 is quite nuts when you can buy a 71 to 86 GHz band, 10 Gbps full duplex radio bridge for under $6000 today. And it'll likely not completely collapse and fail to link at 1.2 km in moderate rain.
oh_my_goodness
The internet is already over lasers, but I guess this is a free space thing.
seanp2k2
When I came to Silicon Valley in the early 2010s and realized that all of the king's horses and all of the king's men could not defeat the scourge that is the US ISP oligopoly even in their own backyards, I knew it was pointless to ever hope for real FTTH in Bay Area markets in the next couple of decades.
Wireless solutions have orders of magnitude less bandwidth than fiber, and you can run lots of fiber in a bundle, whereas there's limited spectrum and only one already-saturated RF environment going wireless.
Only in the past few months has Comcast (the only actual high-speed internet option in our fairly typical Silicon Valley suburban neighborhood of ~$2.5m average homes) deigned to offer upload speeds greater than the previous 35mbit cap...now we can push 200mbit for $120/mo (for 1.2tb monthly combined U+D, add $20 for "unlimited") and you usually have to buy a new modem even if your old one supported the tech, since they only support specific firmware on specific SKUs. Meanwhile, GFiber is offering 8gbps symmetrical for $150/mo unmetered.
Reminder also that Ricochet was wireless internet in 1994 on 900mhz ISM using FHSS in the Valley. Ooh how far we've come in 30 years.
ecshafer
In Silicon Valley, in a neighborhood of $2.5M homes, and thus household incomes on average of probably $600k+. It should be possible to get a fiber rollout, they aren't that expensive. Make a neighborhood ISP, or try and get a municipal ISP.
hn_acc1
I seem to recall a few different "if we get this many signed up" flyers (more East Bay than South) trying to organize something along those lines - maybe the name was Compass? Also try Sonic and see if they have anything for you.
I've had ATT gigabit fiber for ~6(?) years now. And for the first time ever (started with a 300 baud modem in 1985), similar to recent CPU releases, don't feel a compelling need to upgrade. In fact, would consider 0.5 Gbps if it offered substantial savings, since we don't fully use 1 GB. But no caps, and getting measured throughput of 850+ MB/s for $85 (started at $70) is affordable (for SV) and has been very reliable other than when the very-early transceiver they installed (I was one of the first to get it) got waterlogged and shorted out - the replacement was installed indoors, so that won't happen again.
I'm not a fan of ATT's corporate policies by any means. But the alternatives are, well.. Comcast with a much worse service (not symmetrical) or some flavor of 5G.
rcpt
Santa Cruz has one.
walrus01
People want their nice neighborhood of 2.5 m homes where 100% of utilities are underground and aesthetically pleasing. While people in ordinary neighborhoods in Japan with 100% aerial fiber on utility poles have far superior fiber based last mile ISP service.
jerlam
Silicon Valley is an aberration due to its low housing density. Places with more apartments and condos are much more likely to have infrastructure improvements since you can reach a lot more customers with a lot less money.
I live a little north of you, and had Webpass (wireless) in 2012 - $50/mo for 100 Mbps. Now it's $70/mo for 1 Gbps, but I'm with another wireless ISP for $35/mo for 500 Mbps.
yonran
Is this optical phased array beam steering hardware supposed to be affordable for ISPs such as Monkeybrains to use for end users, or is this just for backhaul connections?
mmaunder
This doesn't work in fog. Space lasers are great. No fog. Everyone wants the 500 Terahertz frequencies to work because bandwidth. They have about 25,000 times the carrying capacity of say 20 Ghz. The lower Ghz stuff penetrates weather to varying degrees. Visible light not so much unless it's a vacuum which is perfect. They should move their nodes into space. Oh wait someone already did that.
fweimer
What's so innovative about this? The distance? Some sort of mesh routing? Point-to-point optical wireless links have been available commercially for quite a while.
wmf
They're claiming 10x longer distance than other FSO links I've seen.
xnx
The second episode of the Google Moonshot podcast covers this: https://www.youtube.com/watch?v=RLaYGw5_PE0
I'm low key afraid that this stuff is gonna get popular for .mil usage.
Line of sight free space optics can be immune to many many forms of jamming. Its usage dots the sci books I've read over the years, but almost always for scary reasons.
Here's the Navy today announcing work on AirBorne System for Optical Relay and Broadcast (ABSORB), a (for now) low-cost prototype one-to-many (I maybe mis-inferring what multi-access means?) relayable free space system, https://defence-blog.com/us-navy-plans-to-revolutionize-nava...