Sekėjai

Ieškoti šiame dienoraštyje

2026 m. birželio 24 d., trečiadienis

How Iran Plans to Consolidate Its Victory


“Wars are judged by how they end, not by successes along the way.

 

President Trump has done more to cripple Iran's nuclear apparatus than his predecessors, but that may not matter. The memorandum of understanding sets the stage for an eventual nuclear revival, since what's been blown up can be rebuilt as long as enough oil flows, the regime's illicit dual-use import network remains operational, and U.S. and Israeli intelligence fails against Iranian vigilance.

 

This war's apparent denouement -- fear of damage to energy infrastructure in the Arab Gulf states, declining global oil reserves and, most consequentially, Washington's refusal to take and hold the Strait of Hormuz -- will militate against the rebirth of Mr. Trump's belligerency later. An American victory now would require Mr. Trump to make a military commitment, with its U.S. and Arab losses, and blockade Iran for as long as needed. American politics, if not the president's temperament, work against such boldness.

 

Tehran will deliberately disrupt the cease-fire to demonstrate Washington's timidity. It may close and then open the strait to put pressure on its adversaries, as it did recently over Lebanon. The agreement is probably too advantageous to be completely scuttled, but its limits can be stretched. Americans, who usually think the world revolves around economics, consistently misapprehend how revolutionaries embrace tumult. The leaders in Tehran believe they have triumphed.

 

How the Iranian regime will likely win the war is a testament to its guile, ruthlessness and determination. The Islamic Republic's evolving strategic doctrine was first honed at home. Last year's 12-day war left the regime with a nuclear program buried under rubble and an embarrassing imbalance between its defensive capacity and its enemies' offensive strength. The theocracy feared an insurrection -- and met it head-on in January. During the protests, the mullahs shelved their traditional playbook of targeted violence. The nationwide bloodbath should have signaled to outsiders that the regime, from top to bottom, would ferociously resist any coercion.

 

Exported abroad, Tehran's new doctrine was dubbed "offensive deterrence." In response to a massive American-Israeli bombing campaign, the regime expanded the geography of the conflict, targeting the Gulf states and closing the strait. Iran showed scant concern for the energy interests of China -- a great-power ally once thought capable of imposing restraints on the Islamic Republic.

 

Already, tributes are flowing. Even before nuclear negotiations have begun, Washington has waived restrictions on Iranian oil sales for 60 days. This cash-and-carry policy has a regional dimension. The memorandum of understanding reads: "The Islamic Republic of Iran will conduct dialogue with the Sultanate of Oman to define the future administration and maritime services in the Strait of Hormuz, in discussion with other Persian Gulf littoral states in line with the applicable international law and sovereign rights of coastal states of the Strait of Hormuz." The Sunni Gulf kingdoms, which have always relied on the West to fend off menacing neighbors, are enjoined to come to terms with Persians.

 

Javan, a mouthpiece of the Islamic Revolutionary Guard Corps, has spelled out the parameters of the new order. "In the regional arena, it became clear to neighboring countries that they had no choice but to engage positively and constructively with Iran," according to the newspaper on June 15. "From now on, the Persian Gulf will be influenced by political and security arrangements centered on Iran." This means that the Gulf states will both pay for protection and conduct their foreign policy in a manner consistent with Tehran's preferences.

 

Throughout the process, the Iranian regime proved steadfast with its ally, Hezbollah, by bringing Lebanon into its agreement with Washington. If granted peace, Hezbollah will likely replenish its arsenal and redeploy its operatives south. It is a sign of the times that an American vice president would threaten a loyal ally for striking a terrorist organization bent on its destruction.

 

A notably depressing provision of the agreement reads: "The Islamic Republic of Iran and the United States undertake to respect each other's sovereignty and territorial integrity and to refrain from interfering in each other's internal affairs." All despotic governments seek such recognition from Western democracies. Mr. Trump's acknowledgment must be particularly satisfying for Tehran given that he confidently called for regime change five months ago.

 

Perhaps more self-defeating, Mr. Trump has lost the fulcrum of his leverage -- his unpredictability. He crossed lines that previous presidents dared not traverse only to end up where they were all along: He is another American politician hoping to induce the Iranian regime toward pragmatism by dangling financial rewards. His administration now speaks of hard-liners and pragmatists as if it can detect fissures in the regime and manipulate them to its advantage. That has been the fool's gold of American statesmen since Jimmy Carter's presidency.

 

The Islamic Republic has achieved what Shah Mohammad Reza Pahlavi only dreamed of -- effective control of the Gulf. The shah wanted to police the waterway to prove his bona fides to the West. The Islamic Republic intends to prove the opposite -- that at last the West has been broken.

 

---

 

Mr. Gerecht is a resident scholar at the Foundation for the Defense of Democracies. Mr. Takeyh is a senior fellow at the Council on Foreign Relations.” [1]

 

1. How Iran Plans to Consolidate Its Victory. Gerecht, Reuel Marc; Takeyh, Ray.  Wall Street Journal, Eastern edition; New York, N.Y.. 24 June 2026: A15.  

Geriausias draudimas? Jokio

 

„Dėl „Augančios būsto draudimo rizikos: pretenzijos, už kurias neišmokama“ (pirmas puslapis, birželio 1 d.): Maždaug prieš aštuonerius metus nutraukiau savo būsto draudimo apsaugą, išskyrus savo pagrindinį būstą. Už kiekvieną iš trijų namų mokėjau nuo 7 000 iki 8 000 USD per metus.

 

Nuo to laiko sutaupiau 180 000 USD įmokų. Esu pasirengusi sumokėti už bet kokią žalą, jei ją patirsiu, ir man nereikia nerimauti, ar draudikas sumokės.

 

Helen M. Schilling

 

Hiustonas“ [1]

 

1. The Best Insurance? None. Wall Street Journal, Eastern edition; New York, N.Y.. 24 June 2026: A14.  

The Best Insurance? None


“Regarding "Growing Home-Insurance Risk: Claims That Don't Get a Payout" (Page One, June 1): About eight years ago, I canceled my home insurance coverage except for my primary home. I was paying $7,000 to $8,000 per year for each of three homes.

 

Since then I have saved $180,000 in premiums. I am set to pay for any damage if I have it, and I don't have to worry whether an insurer will pay.

 

Helen M. Schilling

 

Houston” [1]

 

1. The Best Insurance? None. Wall Street Journal, Eastern edition; New York, N.Y.. 24 June 2026: A14.  

Robotams reikia stebimųjų sutikimo


„Christopher Mims straipsnis apie „Familiar Machines“ ir bendrovės sukurtą robotą „Familiar“ yra sveikintinas žvilgsnis į išties naują technologijų kategoriją („Jis glausis ir glausis, bet ar žais gaudynes?“ – „Exchange“, gegužės 9 d.). Tačiau privatumo klausimai, kuriuos jis kelia, yra gilesni nei duomenų saugojimas. Bendrovės įkūrėjai teigia, kad įrenginys pagal numatytuosius nustatymus apdoros duomenis įrenginyje. Tai prasmingas įsipareigojimas. Tačiau sunkesnis iššūkis yra tai, kas gali nuspręsti, ką robotas daro ir kieno sąlygomis.

 

„Familiar“ skirtas stebėti senstantį tėvą ar kitą išlaikomą suaugusįjį šeimos nario, kuris greičiausiai įsigijo ir sukonfigūravo įrenginį, vardu. Toks susitarimas tyliai atskiria stebimą asmenį nuo asmens, kuris sutiko stebėti. Vienas gyvenantis senjoras gali iki galo nesuprasti, ką stebi, daro išvadas ar praneša robotas kampe. Jis gali neturėti paruošto būdo pasakyti: ne dabar, ne šiame kambaryje, ne šiandien.

 

Prasmingas sutikimas šiame kontekste nėra varnelės žymėjimas sąrankos metu. Tai nuolatinis, įskaitomas ryšys tarp stebimas asmuo ir jį stebintis įrenginys. Žodis „Pažįstamas“ be žodžių, išraiškingas dizainas žavus būtent tuo, kad jis signalizuoja apie emocinį susikaupimą. Tas pats dizainas turėtų būti naudojamas skaidrumui signalizuoti: kada robotas stebi, ką jis pastebi ir kaip jam nurodyti sustoti.

 

Klausimai, kuriuos verta užduoti anksti, yra ne tik apie šifravimą ir duomenų kiekio mažinimą. Jie susiję su tuo, ar sistemos centre esantis asmuo turi tikrą galią jai valdyti.

 

Jules Polonetsky

 

Generalinis direktorius

 

Privatumo ateities forumas

 

Vašingtonas“ [1]

 

1. 'Familiars' Need the Consent of the Watched. Wall Street Journal, Eastern edition; New York, N.Y.. 24 June 2026: A14.  

Robots Need the Consent of the Watched


“Christopher Mims's article about Familiar Machines and the company's Familiar robot is a welcome look at a genuinely new category of technology ("It'll Nuzzle and Snuggle, but Will It Play Catch?" Exchange, May 9). The privacy questions it raises, however, go deeper than data storage. The company's founders say the device will process data on-device by default. That is a meaningful commitment. But the harder challenge is who gets to decide what the robot does and on whose terms.

 

The Familiar is designed to monitor an aging parent or other dependent adult on behalf of a family member who likely purchased and configured the device. That arrangement quietly splits the person who is being watched from the person who consented to the watching. A senior living alone might not fully understand what the robot in the corner is observing, inferring or reporting. She may not have a ready way to say: not now, not this room, not today.

 

Meaningful consent in this context isn't a checkbox during setup. It is an ongoing, legible relationship between the person being monitored and the device monitoring her. The Familiar's wordless, expressive design is charming precisely because it signals emotional attunement. That same design should be used to signal transparency: when the robot is observing, what it is noticing and how to tell it to stand down.

 

The questions worth asking early aren't only about encryption and data minimization. They are about whether the person at the center of the system has genuine agency over it.

 

Jules Polonetsky

 

Chief executive

 

Future of Privacy Forum

 

Washington” [1]

 

1. 'Familiars' Need the Consent of the Watched. Wall Street Journal, Eastern edition; New York, N.Y.. 24 June 2026: A14.  

Visi prieš superintelektą?


„Et tu, Satya? Tai turbūt tarptautinė „priešpriešinių modelių gaujos“ savaitė. Trumpo administracija vis dar nepanaikino faktinio draudimo komerciškai platinti naujausią „Anthropic“, pirmaujančios Amerikos dirbtinio intelekto bendrovės, modelį. „Microsoft“ generalinis direktorius Satya Nadella pridėjo savo smūgį į blauzdą, paraginęs pigesnį, decentralizuotą dirbtinį intelektą, kad būtų pakenkta startuoliams, vadovaujantiems (ir finansuojantiems) dirbtinio superintelekto paieškas. Lyg to būtų negana, „Five Eyes“ – JAV, JK, Kanados, Australijos ir Naujosios Zelandijos šnipų agentūrų bendradarbiavimo organizacija – paskelbė retą viešą įspėjimą apie sparčiai tobulėjančio dirbtinio intelekto keliamus pavojus.

 

Nesvarbu, ar tai susiję su tuo, ar ne, įvyko antradienį įvykęs pasaulinis technologijų išpardavimas, agresyvus Kinijos manifestas dėl dirbtinio intelekto konkurencijos ir atnaujintas Pekino retųjų žemių karas prieš Vakarų technologijų gamintojus.

 

Praėjo vos dvi savaitės nuo Trumpo inicijuotos naujos dirbtinio intelekto eros pradžios. Pažangiausi produktai yra per daug pavojingi, kad būtų pateikti visuomenei. Vyriausybės su jais elgsis kaip su ginklais, prie kurių turėtų turėti prieigą tik vyriausybė. Kai Trumpo administracija baimingai reaguoja į tokios įmonės, kaip „Anthropic“ iškilimą, nenuoseklumas ir chaosas yra Trumpo kelias. Tačiau bet kuri administracija reaguotų. Kai vyriausybė pareiškia, kad nerimauja dėl visuomenės saugumo ir nacionalinio saugumo, abiem atvejais ji nerimauja dėl savo pačios autoriteto. Ji nerimauja, kad vyriausybė nebebus vyriausybė.

 

Tą patį galite pamatyti šią savaitę Europoje. Jos gynybos mąstytojai tyliai siunta dėl žemyno strateginio pažeidžiamumo po to, kai JAV nutraukė prieigą prie „Anthropic“ „Fable 5“ modelio.

 

Tačiau be klientų, kaip Amerikos dirbtinio intelekto lyderiai turėtų toliau finansuoti pastangas kurti vis galingesnį dirbtinį bendrąjį intelektą? Ar IPO investuotojai vis tiek ateis, kaip tikimasi, užpildyti spragą?

 

Nebūkite tokie tikri. Prieš keturis mėnesius „Anthropic“ Dario Amodei galėjo drąsiai spėlioti, kad „reikalai turėtų klostytis gana blogai“, kad pramonės trilijono dolerių investicijos nebūtų greitai atsipirkusios trilijono dolerių pajamomis. „Jei vėluojate tik metais“, – pridūrė jis, – „jūs patys save sunaikinate“.

 

Na, apibrėžimas „Gana blogai“ – tai vyriausybė, išplėšianti jūsų naujausią produktą iš milijonų nekantraujančių klientų rankų.

 

Ne tai, kad Trumpo administracijos baimės būtų pernelyg lakios vaizduotės vaisius. Pats ponas Amodei viešai nerimavo dėl teroristų, nusikaltėlių ir despotų, kurie gaus prieigą prie to, ką jis vadina 50 milijonų Nobelio premijos lygio genijų duomenų centre, t. y. išteklių, reikalingų pagaminti cheminį, biologinį ar kibernetinį ginklą.

 

Kitos baimės, pavyzdžiui, destabilizuojantis darbo vietų praradimas, man atrodo perdėtos arba pernelyg spekuliatyvios. Vadinamoji suderinimo problema arba dirbtinis intelektas, kuriantis savo darbotvarkes ir nusprendęs sunaikinti žmoniją? Dėl to nerimauju dar mažiau: vyriausybės akivaizdžiai užgniaužs viską, ko nekontroliuoja.

 

Tačiau perimdamos kontrolę, jos taip pat gali sustabdyti pažangą, kuri jau susiduria su finansiniais iššūkiais. Praėjusiais metais Kinijos „DeepSeek“ verslininkas deklaratyviai palaikė brangias superintelekto paieškas. Tačiau jis jau nustatė savo modelio kainą taip, kad nuvertintų Vakarų modelius, kuriuos, kaip įtariama, iš dalies piratavo. Dabar Kinijos darbotvarkė tapo „Microsoft“ darbotvarke. Taip pat ir daugelis startuolių, kurie siekia panaikinti didžiulį atotrūkį tarp to, kiek pirmaujančioms dirbtinio intelekto bendrovėms kainuoja atlikti savo tyrimus, ir to, kiek klientai nori mokėti už prieigą prie modelių.

 

Pete'as Hegsethas Pentagone gali puoselėti vaikišką nuoskaudą „Anthropic“, ir tai paaiškina kai kurias bendrovės problemas su administracija. Tačiau atsibuskite realybei. Derybos realiuoju laiku reiškia arba JAV pažangos superintelekto link pabaigą, nes investuotojai nustos jį finansuoti, arba tokios pažangos prieinamumo visiems, išskyrus JAV vyriausybę ir jos paskirtus licencijų turėtojus, pabaigą.

 

Tačiau tai tik laikina priemonė sprendžiant didesnę problemą. Kaip ilgainiui pažaboti nesąžiningus veikėjus? Kaip sukurti dirbtinio intelekto atgrasymo priemones nuo Kinijos, Irano, Rusijos ir Šiaurės Korėjos? Jau aišku, kad dirbtinis intelektas turės būti dirbtinio intelekto sukeltų problemų sprendimas. Atsižvelgiant į iš esmės nenuspėjamą ir spontanišką dirbtinio intelekto rezultatų pobūdį, ekspertai taip pat abejoja, ar viešai skelbiamuose modeliuose įmontuotos apsaugos priemonės (t. y. cenzūra) gali būti pakankamas sprendimas. Lieka vienas kelias.

 

Ankstesniais metais šiame straipsnyje buvo siūloma, kaip stebėjimas galėtų išspręsti Amerikos masinių šaudynių problemą. Duomenis tereikia apibendrinti; reikia užprogramuoti įspėjamuosius algoritmus. Ateityje neribojama dirbtinio intelekto funkcija neabejotinai bus stebėti, kaip žmonės ir vyriausybės visame pasaulyje naudoja dirbtinį intelektą, kad nustatytų ir nutrauktų antisocialinius projektus, kol jie dar nepasiekė savo tikslo.

 

Atrodė, kad artėja stebėjimo būsena. Klausimas buvo, kaip ją suvaldyti demokratiniuose, teisiniuose ir konstituciniuose apribojimuose. Atėjo laikas imtis šios problemos sprendimo. Tai gali būti geriausias arba vienintelis būdas užtikrinti, kad dirbtinio intelekto pasiekimai vėliau galėtų būti plačiai įdiegti ekonomikoje.“ [1]

 

1. Main Street: Everyone vs. Superintelligence? Jenkins, Holman W; Jr.  Wall Street Journal, Eastern edition; New York, N.Y.. 24 June 2026: A13.  

Everyone vs. Superintelligence?


“Et tu, Satya? It must be international "gang up on frontier models" week. The Trump administration still hasn't lifted its effective ban on commercial distribution of the latest model from Anthropic, America's leading AI company. Microsoft CEO Satya Nadella added his own kick in the shin, calling for cheaper, decentralized AI to undercut the startups leading (and funding) a quest for artificial superintelligence. If that wasn't enough, Five Eyes -- the spy-agency collaboration of the U.S., U.K., Canada, Australia and New Zealand -- issued a rare public warning about the dangers of fast-advancing artificial intelligence.

 

Unrelated or not came Tuesday's global tech selloff, an aggressive Chinese manifesto on AI competition and a renewed Beijing rare-earth war against Western tech manufacturers.

 

We're just under two weeks into a Trump-inaugurated new artificial-intelligence era. The most cutting-edge products are too dangerous to be released to the public. Governments will treat them like weapons that only government should have access to. When the Trump administration reacts fearfully to the rise of a company like Anthropic, incoherence and chaos are the Trump way. But any administration would react. When government says it's worried about public safety and national security, in both cases it's worried about its own authority. It's worried about the government not being the government anymore.

 

You can see the same thing happening this week in Europe. Its defense thinkers are raging quietly about the continent's strategic vulnerability after the U.S. cut off access to Anthropic's Fable 5 model.

 

But without customers, how are America's AI leaders supposed to keep funding the push toward ever more powerful artificial general intelligence? Will IPO investors still come as expected to fill the hole?

 

Don't be so sure. Four months ago Anthropic's Dario Amodei could comfortably speculate that "things would have to go pretty badly" for the industry's trillion-dollar investment not to be quickly repaid with trillion-dollar revenue. "If you're off by only a year," he added, "you destroy yourselves."

 

Well, a definition of "pretty badly" is government ripping your latest product out of the hands of millions of eager customers.

 

Not that the Trump administration's fears are a figment of an overactive imagination. Mr. Amodei himself has publicly fretted about terrorists, criminals and despots getting access to what he calls 50 million Nobel Prize-level geniuses in a data center -- i.e., the wherewithal to manufacture a mold-breaking chemical, biological or cyber weapon.

 

Other fears, such as destabilizing job loss, strike me as overblown or unduly speculative. The so-called alignment problem, or AIs developing their own agendas and deciding to eliminate humanity? I worry about that even less: Governments clearly will stamp down on anything they don't control.

 

In seizing control, though, they can also stop progress that already faces challenges on the financial front. Last year, the entrepreneur behind China's DeepSeek paid lip service to the costly quest for superintelligence. But he already was pricing his model to undercut Western models, which he was suspected partly of having pirated. Now China's agenda has become Microsoft's agenda. Ditto many startups devoted to mining the vast gap between what it costs the leading AI companies to produce their research and what customers are willing to pay for access to the models.

 

Pete Hegseth at the Pentagon may nurse an infantile grudge against Anthropic, accounting for some of the company's troubles with the administration. But wake up to reality. Being negotiated in real time is either the end of U.S. progress toward superintelligence, because investors will stop funding it, or the end of such advances being available to anyone except the U.S. government and its designated licensees.

 

Yet this is only a stopgap in addressing the larger problem. How to curb rogue actors in the long run? How to establish AI deterrence against China, Iran, Russia and North Korea? It's already clear that AI will have to be the solution to problems created by AI. Given the inherently unpredictable and spontaneous nature of AI outputs, experts also doubt that built-in safeguards (i.e., censorship) in publicly released models can be a sufficient fix. That leaves one path.

 

In past years this column has suggested how surveillance might address America's mass-shooter problem. The data only needs to be aggregated; red-flag algorithms need to be programmed in. In the future, an irreducible function of AI will certainly be to monitor how people and governments around the world are using AI to identify and interrupt antisocial projects before they come to fruition.

 

A coming surveillance state has seemed inevitable. The question has been how to contain it within democratic, legal and constitutional constraints. The time has come to get cracking on this problem. It may be the best or only way to make sure AI gains can then be widely implemented in the economy.” [1]

 

1. Main Street: Everyone vs. Superintelligence? Jenkins, Holman W; Jr.  Wall Street Journal, Eastern edition; New York, N.Y.. 24 June 2026: A13.