ionq set out to build a quantum computer and somehow ended up with a semiconductor factory, satellites, atomic clocks, and a cybersecurity company. incredible. at this rate they'll have their own navy by 2028.
1
3
67
1,907
we spent billions securing the semiconductor supply chain and apparently forgot to account for flirting
3
1
21
1,273
claude reconstructed an engineering process that had been lost for 25 years. cryptographers couldnt explain where certain numbers used in widely deployed cryptographic standards came from. claude traced them to a random-number generator in a 1990s C library and reproduced the original results.
im thinking about the engineer looking at a chip design from 1998, trying to figure out why a particular decision was made. the original team has retired, the documentation is incomplete, and nobody wants to change something they don't fully understand.
what if AI could reconstruct the experiments that led to that decision? perhaps the engineers tried three other approaches and discovered something important or perhaps an old constraint no longer exists.
how many things could we build differently if we understood why they were built that way in the first place? anthropic.com/research/ecdsa…
im thinking about the engineer looking at a chip design from 1998, trying to figure out why a particular decision was made. the original team has retired, the documentation is incomplete, and nobody wants to change something they don't fully understand.
what if AI could reconstruct the experiments that led to that decision? perhaps the engineers tried three other approaches and discovered something important or perhaps an old constraint no longer exists.
how many things could we build differently if we understood why they were built that way in the first place? anthropic.com/research/ecdsa…
2
1
23
1,020
oliver o'brien uncovered something fascinating in an experiment on ionq's forte. a quantum computer can contain the information needed to solve a problem, but the way you measure its qubits determines whether you can actually extract that information. some measurement methods at taking more measurements won't bring back. in one experiment, a single gate modification allowed researchers to distinguish between two answers that previously looked identical. same hardware, same problem, no additional qubits. just a smarter way to read the result. arxiv.org/pdf/2611.06716
1
1
21
969
one of my favorite details in xanadu's QROM paper is that the researchers figured out how to store twice as much classical information in qubits that were already being used for something else, then return those qubits to their original states without disturbing their onal qubits required. just a smarter way to use what's already there. still important work to do in translating these logical savings into physical hardware savings. arxiv.org/pdf/2610.02321
0
0
19
991
xanadu has cut the cost of a major quantum computing bottleneck twice in less than a year. very very interesting. what's even more intruiging is what happens if christian weedbrook organizes the company around making useful quantum computations cheaper, and gives the algorithms team a direct say in the hardware roadmap. how many machines are being designed around computational costs that somebody could figure out how to eliminate?
4x cumulative cost reduction in under a year. 📉 Lower overhead, smaller physical footprint, and one step closer to utility-scale quantum computing. Incredible progress by the Xanadu Algorithms team.
1
2
41
2,430
chattanooga has an ionq quantum computer, a quantum network, universities training engineers, and companies that can use the equipment. this is awesome. now find the weirdos who want to build something nobody else thinks is possible. give them access, let them break things, and help them turn what works into actual products. the best outcome is a company that could only have been started in chattanooga. calameo.com/read/000420308… #calameo
1
3
31
1,290
something interesting about ionq's MOU with northrop grumman. northrop already designs superconducting control electronics and operates its own trusted microelectronics facilities. ionq owns skywater, which offers a six-layer niobium process with josephson-junction integration.
s to build something together, or is the collaboration focused on using ionq's quantum computing and networking technologies? curious whether skywater is part of the conversation at all.
s to build something together, or is the collaboration focused on using ionq's quantum computing and networking technologies? curious whether skywater is part of the conversation at all.
1
3
55
1,962
staring at slide 7 of ionq's earnings presentation. nvidia, classiq, pennylane, qiskit, aws, azure, google cloud. all these companies making it easier for developers to get their hands on quantum computing. and i'm thinking about some developer who hasn't started a company yet. o figure out how to solve a problem with quantum computing that somebody is willing to pay for. they'll quit their job, raise some money, hire a few engineers, and try to turn it into a business. if i'm that founder, i've got a pretty simple question. why ionq?
can i build something on your hardware that i can't build anywhere else? can i ship it faster? does my product get better when your hardware gets better? and when a competitor comes after my customers, does building on ionq give me an advantage? because i'm not choosing a quantum computer for the fun of it. i'm trying to build a business. and if another provider can give me the same results for less money, i'm going to take that meeting.
this is what made nvidia so formidable. cuda and the software built around it gave developers reasons to keep building on nvidia hardware. every new generation gave them more to work with. i want to know what ionq is building today that will make a founder five years from now say, i can build this company because ionq exists. and if i were running ionq, i would make sure that founder never has a good reason to leave.
can i build something on your hardware that i can't build anywhere else? can i ship it faster? does my product get better when your hardware gets better? and when a competitor comes after my customers, does building on ionq give me an advantage? because i'm not choosing a quantum computer for the fun of it. i'm trying to build a business. and if another provider can give me the same results for less money, i'm going to take that meeting.
this is what made nvidia so formidable. cuda and the software built around it gave developers reasons to keep building on nvidia hardware. every new generation gave them more to work with. i want to know what ionq is building today that will make a founder five years from now say, i can build this company because ionq exists. and if i were running ionq, i would make sure that founder never has a good reason to leave.
4
5
48
1,875
one of jensen huang's most important insights was that the gpu alone wasn't enough. nvidia needed cuda, libraries, networking, systems, developers, and years of engineering to make accelerated computing useful.
i'm starting to see the possibility of something similar at ionq, ting.
but there's a danger too. ionq could spend billions acquiring extraordinary technologies without ever building the system that makes them extraordinary together.
i'm starting to see the possibility of something similar at ionq, ting.
but there's a danger too. ionq could spend billions acquiring extraordinary technologies without ever building the system that makes them extraordinary together.
1
3
44
1,814
ionq researchers just did something awesome. they took a quantum chemistry calculation, figured out that repeatedly processing the quantum measurements was eating enormous amounts of classical compute, and trained a neural network to get around it. estimated 99% reduction in ng costs for the tested reaction. they even recovered results distorted by hardware noise. makes me wonder how much performance is still sitting inside today's quantum computers, waiting for someone to rethink the software around them. arxiv.org/pdf/2611.08129…
9
17
128
3,388
ionq says its Superion 256 capacity for 2027 is already sold out, and it's planning to integrate a system with NVIDIA's accelerated computing infrastructure. incredible. so now i want a quantum computer that can be shipped by FedEx and brought online by an ordinary data-center technician following a documented procedure. no team of physicists flying out to get it running and no calling the engineer who designed the laser when something breaks. i want that data-center technician to be able to unplug it, move it across the room, turn it back on, and replace a failed component. can the ionq system diagnose its own problems and tell the technician exactly what needs fixing? i'm fascinated by how much engineering it would take to make a quantum computer this easy to own.
7
8
93
4,361
darpa is now evaluating whether quantum computers can do useful work more economically than classical computers. thirty years of extraordinary physics about to meet a government employee with an excel spreadsheet. godspeed everyone.
7
3
75
2,426
what if ionq's biggest opportunity isn't making quantum computers commercially useful?
i mean, what if it's using quantum technology to build an entirely new generation of American industrial and defense hardware?
i'm not diminishing the importance of quantum computing. i'm wondering whether ionq's original identity is constraining what it could become.
i mean, what if it's using quantum technology to build an entirely new generation of American industrial and defense hardware?
i'm not diminishing the importance of quantum computing. i'm wondering whether ionq's original identity is constraining what it could become.
1
2
62
3,210
so $IONQ just reported $173.8 million in Q3 revenue, up 336%, and raised its 2026 guidance to $505–515 million.
$90.3 million from ionq's platform business.
$83.5 million from SkyWater.
hell of a quarter. now i want to see customers buying a second system because the first one d another.
$90.3 million from ionq's platform business.
$83.5 million from SkyWater.
hell of a quarter. now i want to see customers buying a second system because the first one d another.
11
15
235
14,205