Quan­tum rev­o­lu­tion

Chad Rigetti prom­ises to de­liver an en­tire new UK tech ecosys­tem

The Daily Telegraph - Business - - Front Page - Harry de Quet­teville

There are some tech­nolo­gies that prom­ise so much that we re­turn to them time and again, de­spite myr­iad fail­ures. Like se­ri­ally re­buffed suit­ors, re­searchers into nu­clear fu­sion or ar­ti­fi­cial gen­eral in­tel­li­gence (AGI) keep pick­ing them­selves up, hop­ing that one day, all their dreams will be­come true.

Some­times it takes a decade af­ter a par­tic­u­larly big dis­ap­point­ment for op­ti­mism to re­assert it­self. But it al­ways hap­pens. The prospect of mas­ter­ing one of that tiny num­ber of truly world-chang­ing tech­nolo­gies is just too al­lur­ing to give up. This time, ev­ery­one thinks, things will be dif­fer­ent.

So it is with quan­tum com­put­ing. Since the late Sev­en­ties, sci­en­tists have un­der­stood that by har­ness­ing the weird-but-true prop­er­ties of sub­atomic par­ti­cles, like elec­trons, as de­scribed by quan­tum the­ory, a new breed of com­put­ers could be made far more pow­er­ful than tra­di­tional ma­chines that rely on sil­i­con chips.

In 1981, physi­cist Richard Feyn­man gave a speech imag­in­ing just such a com­puter. “It’s not a Tur­ing ma­chine,” he noted, re­fer­ring to the def­i­ni­tion of “clas­si­cal” com­put­ers set out by Alan Tur­ing in 1936. “But a ma­chine of a dif­fer­ent kind.”

Now, af­ter 40 years of try­ing to make prac­ti­cal re­al­ity live up to quan­tum the­ory, Bri­tain is get­ting its first of this dif­fer­ent kind of ma­chine. And the man tasked with de­liv­er­ing it is Chad Rigetti. Rigetti is founder of Rigetti com­put­ers, which has just been an­nounced as the leader of the £10m, three-year project, co-funded by gov­ern­ment and in­dus­try, to build the quan­tum com­puter in Abing­don, Oxon.

The square-jawed, Cana­dian-born physi­cist, now based in Cal­i­for­nia, is evan­gel­i­cal about quan­tum com­put­ing’s po­ten­tial: “I be­lieve it’s an ex­tremely valu­able and im­por­tant tech­nol­ogy to gov­ern­ments and na­tions around the world”. Its rev­o­lu­tion­ary power, he says, makes it a “node tech­nol­ogy” – an en­abler to un­leash dozens of other branches of en­deav­our.

“The coun­tries that re­ally have lead­ing edge ef­forts in mak­ing quan­tum com­put­ing prac­ti­cal give them­selves the seeds of an in­no­va­tion hub akin to Sil­i­con Val­ley,” he says. “It’s crit­i­cal to na­tions that view tech­nol­o­gy­driven eco­nomic devel­op­ment as a core part of their na­tional strat­egy.”

That cer­tainly seems to chime with the Do­minic Cum­mings agenda. In­deed, an­nounc­ing the Rigetti deal, Amanda Sol­loway, the science min­is­ter, de­clared the Gov­ern­ment’s am­bi­tion “to be the world’s first quan­tum-ready econ­omy, which could pro­vide UK busi­nesses and in­dus­tries with bil­lions of pounds worth of op­por­tu­ni­ties”. Yet UK gov­ern­ment in­ter­est in quan­tum did not ar­rive with Mr Cum­mings. Con­certed in­vest­ment has been un­der way since 2014, with the in­cep­tion of the Na­tional Quan­tum Tech­nolo­gies Pro­gramme. Last sum­mer, com­bined public and pri­vate in­vest­ment up to that point reached £1bn – “keep­ing pace”, the Gov­ern­ment in­sisted, “with the US and China”.

Not quite. State fund­ing alone in the US over five years from 2018 is $1.2bn (£900m), but that’s with­out the bil­lions in­vested by lead­ing com­pa­nies, no­tably IBM, Honey­well and Google. And while China’s ini­tial in­vest­ment in its own Na­tional Lab­o­ra­tory for Quan­tum In­for­ma­tion Sciences is just over $1bn, some an­a­lysts ex­pect it to hit more than $10bn in the next decade. World­wide, con­sul­tants Qureca es­ti­mate $22bn has been com­mit­ted to quan­tum re­search. That is a sig­nif­i­cant amount, es­pe­cially for a tech­nol­ogy that has, to put it po­litely, doubters. “It’s re­ally hard,” Rigetti says of build­ing quan­tum com­put­ers. “And re­ally, re­ally ad­vanced tech­nol­ogy, when it’s con­tin­u­ally pi­o­neer­ing, I think, will al­ways face a lot of scep­ti­cism.”

Like a souf­fle, quan­tum spec­u­la­tion when over­done is sub­ject to a hope­hype-flop cy­cle. That can be seen in Rigetti’s val­u­a­tion at var­i­ous stages of fundrais­ing in its seven-year ex­is­tence. Its 2014 seed round, ac­cord­ing to Pitchbook, val­ued it at £15m; just three years later in­vestors took stakes that put its worth at £313m. Last month, how­ever, in­vestors ac­quired 57pc of the com­pany for just £62m, valu­ing it at just £101m. Part of the prob­lem, thinks Rigetti, is that while other “deep tech” prob­lems, like AGI, are eas­ily com­pre­hen­si­ble if tricky to re­alise, quan­tum tech­nol­ogy up­sets our very per­cep­tion of re­al­ity.

In clas­si­cal com­put­ers, a flow of elec­trons cre­ates a cur­rent that switches tran­sis­tors, usu­ally made from sil­i­con, on or off. As­ton­ish­ing ad­vances in en­gi­neer­ing means tens of bil­lions of tran­sis­tors can now be packed on to a sin­gle mi­crochip. But each tran­sis­tor still ex­ists in this state – on or off, one or zero – known as a “bit” (or bi­nary digit), which is the small­est unit of data in a com­puter.

Quan­tum com­put­ers, on the other hand, rely on qubits, where elec­trons ex­ist in “su­per­po­si­tion” – on or off, or some­thing in be­tween. Not only this, but these qubits can be “en­tan­gled” with each other, cre­at­ing ex­po­nen­tial growth in com­put­ing power.

Even Ein­stein was fazed by the weirdness of such phe­nom­ena. He called en­tan­gle­ment, which links par­ti­cles in­stantly over huge dis­tances, sug­gest­ing a faster-than-light con­nec­tion, “spooky ac­tion at a dis­tance”. And one ex­pla­na­tion for su­per­po­si­tion, where par­ti­cles ex­ist in mul­ti­ple states un­til they are ob­served, is that mul­ti­ple states are re­quired for mul­ti­ple uni­verses: a “mul­ti­verse” where mul­ti­ple re­al­i­ties ex­ist con­cur­rently. Such bog­gling con­cepts, Rigetti ad­mits, are a prob­lem for se­ri­ous quan­tum re­search be­cause “there’s still a bit of a view that quan­tum me­chan­ics is magic ver­sus a rig­or­ous math­e­mat­i­cal the­ory that thor­oughly and com­pletely de­scribes how the uni­verse op­er­ates. But it’s not mag­i­cal. It works, like New­ton’s laws work. And we don’t find those par­tic­u­larly mag­i­cal any­more. Though I’m sure at one point, the no­tion that there is an at­trac­tive force be­tween the moon and the Earth did feel quite mag­i­cal.”

Still, there are many math­e­ma­ti­cians un­fazed by the com­plex­i­ties who still be­lieve use­ful quan­tum com­put­ers are not so much mag­i­cal, as an il­lu­sion.

The trou­ble is that con­trol­ling qubits is ex­tremely hard. It is not just about the flow of elec­trons, as in sil­i­con tran­sis­tors, but their spin. And quan­tum states are frag­ile, prone to be knocked-off true by en­vi­ron­men­tal “noise” – de­liv­er­ing er­ror-laden re­sults. Sta­bil­is­ing qubits and cor­rect­ing for er­rors are two ma­jor chal­lenges.

Whether these can ever be over­come to en­able com­put­ers pow­ered by chains of mil­lions of en­tan­gled qubits is open to ques­tion. IBM’s cur­rent cham­pion has just 65 qubits, though this week it promised a 1,000 qubit ma­chine by 2023. By the end of the decade it prom­ises to push that to one mil­lion.

Rigetti says his own com­pany has “built real hard­ware, real com­put­ers, that are true gate model quan­tum com­put­ers” and has pro­gressed to an er­ror-prone if promis­ing state known in the busi­ness as“noisy intermedia­te scale quan­tum com­put­ing ”.

The fu­ture, he says, has “an im­mense pos­si­bil­ity to de­liver sub­stan­tial end-user value”. By con­trast, an anony­mous poster, claim­ing to be a former “quan­tum en­gi­neer” on the em­ployee re­view site Glass­door, de­scribes work­ing at Rigetti as “a very hum­bling ex­pe­ri­ence in this age of

‘I be­lieve it’s an ex­tremely valu­able and im­por­tant tech­nol­ogy to gov­ern­ments and na­tions around the world’

techno hubris”. “Chad,” the post con­tin­ues, “is quite un­hinged and en­joys the priv­i­lege of in­def­i­nitely sus­pend­ing re­al­ity.”

The UK con­sor­tium that Rigetti leads shows how far there is to go. Its part­ner, Ox­ford In­stru­ments, will sup­ply state of the art re­frig­er­a­tion needed; a start-up called Phase­craft will de­velop the en­tirely new kind of al­go­rithms re­quired to put qubits to work; and the Univer­sity of Ed­in­burgh will be de­vel­op­ing ways to test and ver­ify both this new hard­ware and soft­ware.

Stan­dard Char­tered Bank is the only in­dus­try out­sider: it is look­ing to har­ness quan­tum com­put­ing to get an edge in fi­nance. But the hope, and Rigetti’s busi­ness model, is that count­less such com­pa­nies will ul­ti­mately pay to ac­cess quan­tum com­put­ing’s power, just as they do to­day to ac­cess cloud com­put­ing ser­vices run by Ama­zon, Google, Mi­crosoft and oth­ers.

In the near term, which he de­fines as “three-to-five” years, Rigetti hopes the tech­nol­ogy will make dra­matic im­pact in three prin­ci­pal are­nas: mol­e­cule dis­cov­ery – which could aid the devel­op­ment of new drugs and new ma­te­ri­als – ma­chine learn­ing and process op­ti­mi­sa­tion for or­gan­i­sa­tions “from Net­flix … to the mil­i­tary”.

Each de­pends on spe­cific ad­van­tages that quan­tum of­fers over clas­si­cal com­put­ing: sub­atomic ma­nip­u­la­tion; un­cer­tainty; and brute com­put­ing power re­spec­tively.

Last year, in a very spe­cific, ab­struse, ex­per­i­ment, Google claimed its ma­chine, named Sy­camore, had demon­strated “quan­tum supremacy” by solv­ing a prob­lem in a few sec­onds that would have taken a clas­si­cal su­per­com­puter 10,000 years. And last month, the US tech gi­ant used Sy­camore to do some­thing far more prac­ti­cal: sim­u­late a chem­i­cal re­ac­tion.

It all amounts, says IBM, to “an in­flec­tion point” for quan­tum com­put­ing, “a mile­stone that marks our abil­ity to im­ple­ment er­ror correction and scale up our de­vices”.

Of course, scep­tics will say they have heard it all be­fore. Even Chad Rigetti con­cedes that, frankly, only see­ing will

be

‘Quan­tum com­put­ing thor­oughly tests the laws of quan­tum physics ev­ery time you send an in­struc­tion to it’

be­liev­ing. “Quan­tum com­put­ing thor­oughly tests the laws of quan­tum physics ev­ery time you send an in­struc­tion to it. To build com­put­ers, you have to com­pletely mas­ter that un­der­ly­ing phys­i­cal the­ory. And that’s what is hap­pen­ing. As we do that, that will lead to a de­mys­ti­fi­ca­tion. In the next gen­er­a­tion or two, quan­tum com­put­ing will feel much more nat­u­ral and straight­for­ward.”

In other words the “magic” will be dis­pelled. Or the il­lu­sion ex­posed.

Chad Rigetti is charged with cre­at­ing a Bri­tish quan­tum com­puter

Newspapers in English

Newspapers from UK

© PressReader. All rights reserved.