Artificial Intelligence (A Special Report) --- Yes, AI Can Make Mistakes. AI Can Find Them, Too.: Since chatbots hallucinate their own facts, it's useful (and easy) to have a second, nitpicking AI that can audit the results for errors

“Whenever you use AI as a research assistant, subject-matter expert, or souped-up search engine, you need to grapple with the risk of hallucination -- AI's tendency to make up its own facts. Your first line of defense against these fabrications? More AI.

 

I now make a point of getting AI to check every fact it gives me. It still isn't foolproof, and it is wise to use an actual flesh-and-blood human to fact-check anything you must get right. But using AI for a round of fact-checking can make the human fact-checking process go faster.

 

To start, before I even read AI-generated research, I get another AI to check its accuracy. I might use the same platform that generated my initial report, but I always start a brand-new session; otherwise the logic that influenced the initial report can influence the fact-checking process.

 

I set up this new session by harnessing the same sycophantic, people-pleasing tendencies that can make AI hallucinate in the first place. If the first AI created its own facts to please me, my AI fact-checker needs to please me by finding everywhere the first AI went wrong.

 

To unleash that nitpicking second AI, I start my fact-checking prompts with instructions like "You are a professor of journalism on your university's ethics board, and it's your job to investigate the work of a research team that's been using AI to generate reports." Or "You're an auditor who has been hired to check the work of an internal data-analytics team that's been using AI to compile customer data and sales prospects." You get the idea. I want the second AI to be every employee's nightmare.

 

I next tell my virtual fact-checker to list every so-called fact that needs checking: every name or citation, every data point and every reported finding. I tell it to find a source for every one of these assertions, and if it's reviewing a research memo that includes hyperlinks or article sources, to click through to each of these links and read the original article with its own virtual eyes.

 

This is often how I catch big and small mistakes: My AI fact-checker might report that a supposed source article doesn't exist, or that a quote from the article was taken out of context.

 

I then instruct the fact-checker to build a table that lists all the facts it's checking and mark each one as "true," "false," "ambiguous" or "unsupported." I tell it to add a column for notes (so it can tell me how it reached its conclusions) plus a column for the source, and (this is crucial!) a column where it includes an exact quote from the source, backing up its conclusion. That allows me to fact-check the fact checker by searching the source article to make sure that quote is really there, and that it's been interpreted correctly.

 

Finally, I go one step further: I tell the fact-checking AI to give me a revised version of the research it checked, reflecting all the corrections it found.

 

When I'm using AI to help with especially high-stakes research, I take steps that make these accuracy checks even more effective. Instead of giving my fact-checking prompt to a single imaginary nit-picker, I use two (starting fresh each time). Then I give their results to a third AI fact-checker and ask it to identify anywhere the two previous nit-pickers disagreed and do its own research to cast a tiebreaking vote.

 

If this sounds like a lot of work, or unduly time-consuming, remember that it's still a lot faster than rewinding weeks of work or a mistaken decision based on incorrect information. Or doing the research yourself from the start.

 

You can also set up your AI tools to accelerate the fact-checking process, for example by turning your favorite nitpicking prompt into an AI assistant (like a custom GPT or a Claude Project), so that you can give any first-draft AI research to it and get back a list of checked facts and a revised memo.

 

Or use Claude Code (or Claude Cowork) to spin up a whole team of fact-checkers: Tell the AI that once it has built a list of facts to check, it should assign those facts in small batches to a team of fact-checkers so that each fact gets checked by at least two different AIs, and where needed, a third tiebreaker. When an AI fact-checker hands out this work in batches to its many virtual helpers, even a detailed research brief can get verified (and corrected) quite quickly.

 

Of course, it doesn't end there. Even after I run my AI research briefs through my AI fact-checker, and even if I run those corrected briefs through another entire loop of fact-checking, I sometimes find quirks or inconsistencies that lead me to do further digging, and to read through some of the underlying research with my own human eyes.

 

But that doesn't make all the AI fact-checking pointless. On the contrary: Using AI to catch the most obvious AI mistakes is what gives me the room to dive deep into the research questions that demand human intelligence.

 

---

 

Alexandra Samuel is a technology researcher and host of the AI podcast "Me + Viv." Email her at reports@wsj.com.” [1]

 

1. Artificial Intelligence (A Special Report) --- Yes, AI Can Make Mistakes. AI Can Find Them, Too.: Since chatbots hallucinate their own facts, it's useful (and easy) to have a second, nitpicking AI that can audit the results for errors. Samuel, Alexandra.  Wall Street Journal, Eastern edition; New York, N.Y.. 26 May 2026: R5. 

„Alphabet“ investicijų padalinys remia „OpenRouter“, kuri padeda įmonėms pasirinkti iš šimtų modelių skirtingoms programinės įrangos užduotims.

 

Įmonės vis dažniau naudoja daugiau nei vieną dirbtinio intelekto modelį savo programinei įrangai, siekdamos efektyvumo ir vis labiau dėl kainos. Tai skatina paklausą ir investuotojų susidomėjimą startuoliu, kuris padeda joms tai padaryti.

 

Prieš trejus metus įkurtas startuolis „OpenRouter“ save apibūdina kaip dirbtinio intelekto modelių prekyvietę. Antradienį jis planuoja paskelbti, kad pritraukė 113 mln. USD, vadovaujamas „CapitalG“, „Alphabet“, „Google“ patronuojančios bendrovės, investicijų padalinio.

 

Šis finansavimo etapas vertina „OpenRouter“ apie 1,3 mlrd. USD, teigia asmuo, žinantis apie finansavimo etapą ir neturėjęs įgaliojimų viešai atskleisti šios sumos. Tai daugiau nei dvigubai daugiau nei įmonės vertė ankstesniame lėšų rinkimo etape praėjusiais metais.

 

Alexas Atallahas, „OpenRouter“ įkūrėjas ir generalinis direktorius, savo įmonę apibūdino kaip dirbtinio intelekto atitikmenį finansų milžinei „Stripe“, kuri tvarko visus klientų mokėjimus per vieną prieigos tašką. Tai taip pat padeda įmonėms išvengti prisirišimo prie vieno konkretaus modelio tiekėjo.

 

„Daug ką mes darome, tai bandome padėti žmonėms rasti tinkamą tiekėją darbui“, – sakė jis.

 

Augantis „OpenRouter“ populiarumas – kuris teigia, kad dabar kiekvieną savaitę apdoroja 25 trilijonus žetonų, pagrindinį dirbtinio intelekto skaičiavimo vienetą, palyginti su penkiais trilijonais prieš šešis mėnesius – pabrėžia augantį galimų modelių skaičių ir spartų žetonų ekonomikos augimą. (Ponas Atallah atsisakė pateikti finansinės veiklos statistiką, įskaitant pajamas.)

 

Nors plačioji visuomenė žino kai kuriuos, įskaitant „OpenAI“, „Anthropic“ ir „Google“, „OpenRouter“ teigia, kad suteikia prieigą prie daugiau nei 400, kai kurie teikia pirmenybę sudėtingam mąstymui, o kiti – greičiui.

 

Turbūt ne mažiau svarbu tai, kad įmonės vis labiau suvokia dirbtinio intelekto naudojimo išlaidas, ypač todėl, kad jos labiau pasikliauja agentais, kurie sunaudoja daugiau žetonų ir skatina darbuotojus daugiau naudoti dirbtinį intelektą. „Uber“ technologijų vadovas Praveenas Neppalli Naga neseniai „The Information“ sakė, kad Pavėžėjimo paslaugų milžinė visą savo 2026 m. dirbtinio intelekto biudžetą išleido vos po kelių metų mėnesių.

 

Tai vis dažniau reiškia, kad daugelis įmonių naudoja atvirojo kodo modelius, kuriais galima naudotis nemokamai, kad sumažintų išlaidas. Ponas Atallah teigė, kad „OpenRouter“ padeda klientams gerokai sumažinti išlaidas. Jis pridūrė, kad jei įmonės nebus atsargios, užklausų vykdymas per dirbtinio intelekto modelius, vadinamas išvadomis, gali sukelti „begalinį sąnaudų centrą“.

 

Iš tiesų, populiariausi „OpenRouter“ klientų naudoti modeliai per pastarąją savaitę buvo atvirojo kodo modeliai iš Kinijos bendrovių „DeepSeek“ ir „Tencent“, po jų sekė aukščiausios klasės „Anthropic“ „Claude 4.7 Opus“.

 

Būtent toks požiūris į dirbtinio intelekto bumą, iš esmės kaip vartai į modelius, patraukė „CapitalG“ į „OpenRouter“, teigia investicinės įmonės partneriai Jane Alexander ir Mo Jomaa. Nors kitos įmonės, pavyzdžiui, „Vercel“, taip pat leidžia klientams pasirinkti skirtingus dirbtinio intelekto modelius skirtingoms užduotims, p. Jomaa teigė, kad augantis „OpenRouter“ populiarumas suteikia jai pranašumą mokantis, ko nori klientai, ir nukreipiant juos prie geriausių variantų.

 

P. Atallah teigimu, „OpenRouter“ neieškojo lėšų, kai „CapitalG“ prieš kelis mėnesius išreiškė susidomėjimą investuoti. Tačiau įmonė atsiuntė, jo žodžiais tariant, įspūdingą „atvirkštinio pristatymo rinkinį“, ir jie greitai suplanavo naują lėšų rinkimo etapą.

 

Kiti šio etapo dalyviai yra rizikos kapitalo „Nvidia“, „ServiceNow“, „MongoDB“, „Snowflake“ ir „Databricks“ padaliniai, taip pat esami rėmėjai Andreessenas Horowitzas ir „Menlo Ventures“.

 

Ponas Atallah teigė, kad „OpenRouter“ planuoja panaudoti kapitalą tolesniam augimui, įskaitant daugiau inžinierių samdymą ir daugiau produktų siūlymą.“ [1]

 

1. A One-Stop Shop for A.I. Models Raises $113 Million. Michael J. de la Merced.  New York Times (Online) New York Times Company. May 26, 2026.

A One-Stop Shop for A.I. Models Raises $113 Million

 

 

“An investment arm of Alphabet is backing OpenRouter, which helps companies choose among hundreds of models for different software tasks.

 

Companies increasingly rely on more than one artificial intelligence model to power their software, for efficiency — and increasingly, for cost. That is driving demand, and investor interest, in a start-up that helps them do just that.

 

The start-up, OpenRouter, which was founded three years ago, describes itself as a marketplace for A.I. models. It plans to announce on Tuesday that it has raised $113 million, led by CapitalG, an investment arm of Alphabet, Google’s parent company.

 

The round values OpenRouter at about $1.3 billion, according to a person with knowledge of the round who wasn’t authorized to disclose the figure publicly. That is more than double the company’s value in its previous fund-raising round last year.

 

Alex Atallah, an OpenRouter founder and its chief executive, described his company as the A.I. equivalent of Stripe, the financial giant that handles all payments for customers through a single access point. It also helps companies avoid being locked into one particular model provider.

 

“A lot of what we do is try to help people find the right vendor for the job,” he said.

 

The rising popularity of OpenRouter — which says it now processes 25 trillion tokens, the basic unit of A.I. computing, every week, up from five trillion six months ago — underscores the growing number of models available and the rapid rise of the token economy. (Mr. Atallah declined to provide financial performance statistics, including revenue.)

 

While the general public knows of some, including those from OpenAI, Anthropic and Google, OpenRouter says it provides access to more than 400, with some prioritizing heavy-duty reasoning and others speed.

 

Perhaps just as important, companies are increasingly mindful of the costs of A.I. use, especially as they rely more on agents that consume more tokens and push employees to use A.I. more. The chief technology officer of Uber, Praveen Neppalli Naga, recently told The Information that the ride-hailing giant spent its entire A.I. budget for 2026 just a few months into the year.

 

That increasingly means many companies are using open-source models, which are free to use, to help cut down on costs. Mr. Atallah said that OpenRouter was helping customers significantly shrink their expenses. He added that if companies aren’t careful, running queries through A.I. models, known as inference, could end up leading to “an infinite cost center.”

 

Indeed, the most popular models used by OpenRouter’s customers over the past week were open-source ones from the Chinese companies DeepSeek and Tencent, followed by Anthropic’s high-end Claude 4.7 Opus.

 

It was that approach to the A.I. boom, essentially being a gateway to models, that drew CapitalG to OpenRouter, according to Jane Alexander and Mo Jomaa, partners at the investment firm. While other companies, like Vercel, also let customers choose different A.I. models for different tasks, Mr. Jomaa said that OpenRouter’s rising popularity gave it an edge in learning what customers want and steering them to the best options.

 

OpenRouter had not been looking to raise money when CapitalG expressed interest in investing several months ago, Mr. Atallah said. But the firm sent what he called an impressive “reverse pitch deck,” and they quickly lined up a new fund-raising round.

 

Other participants in the round include the venture capital arms of Nvidia, ServiceNow, MongoDB, Snowflake and Databricks, as well as the existing backers Andreessen Horowitz and Menlo Ventures.

 

Mr. Atallah said that OpenRouter planned to use the capital to continue growing, including by hiring more engineers and offering more products.” [1]

 

1. A One-Stop Shop for A.I. Models Raises $113 Million. Michael J. de la Merced.  New York Times (Online) New York Times Company. May 26, 2026.

„Mirtis Amerikai“: Irano aukščiausiasis lyderis perspėja, kad JAV neturės „saugaus prieglobsčio“ – žada ir Izraelio sunaikinimą

 

„Irano aukščiausiasis lyderis Mojtaba Khamenei, kuris nebuvo matytas ir apie kurį nebuvo girdėti viešai nuo tada, kai po pirmųjų JAV ir Izraelio smūgių Iranui pakeitė savo nužudytą tėvą, antradienį pareiškė, kad „Mirtis Amerikai“ ir „Mirtis Izraeliui“ taps musulmonų pasaulio susibūrimo šūkiais, kartu prisiekdamas, kad žydų valstybė artėja prie „paskutinių savo egzistavimo etapų“.

 

Rašytinėje žinutėje, skirtoje kasmetinei Hadžo piligriminei kelionei į Meką, Khamenei Izraelį pavadino „vėžiniu augliu“, pakartojo savo tėvo prognozę, kad žydų valstybė neišgyvens dar 25 metų, palyginti su prieš dešimtmetį, – faktiškai Izraelio sunaikinimas buvo nukeltas į maždaug 15 metų, – ir perspėjo, kad regioninės tautos nebebus „skydai“ Amerikos karinėms bazėms.

 

„„Mirtis Amerikai“ ir „Mirtis Izraeliui“ taps bendromis giesmėmis islamo ummos [tautai] ir pasaulio engiamiesiems, ypač jaunimui“, – rašė Khamenei pareiškime, paskelbtame su Irano valstybe susijusių žiniasklaidos priemonių ir savo oficialioje X paskyroje, gyrė „Hezbollah“ ir kitas Irano remiamas teroristines grupuotes už tai, ką jis pavadino pergalėmis prieš „amerikiečių sionistų armijas“.

 

Khamenei taip pat paragino musulmonus melstis už „Palestinos ir Al Aksos mečetės išlaisvinimą“ ir už „galutinę pergalę prieš pasaulinę aroganciją“ – terminą, kurį režimas įprastai vartoja apibūdindamas Jungtines Valstijas.

 

„Sukrėstas sionistų režimas ir vėžinis Izraelio auglys artėja prie paskutinių savo apgailėtino egzistavimo etapų“, – kaltino jis, pridurdamas, kad „Jungtinės Valstijos ne tik nebeturės saugaus prieglobsčio savo išdaigoms ir karinių bazių kūrimui regione, bet ir diena iš dienos vis labiau tolsta nuo savo buvusio statuso“.

 

Khamenei toliau gyrė tai, ką jis apibūdino kaip „Allahu Akbar“ galią, teigdamas, kad Irano pajėgos ir sąjungininkų džihadistų grupuotės smogė Jungtinėms Valstijoms „stiprų antausį“, o Izraelis „iškvėpė paskutinį atodūsį“.

 

„Islamiškojo Irano ginkluotosios pajėgos, pasirengusios paaukoti savo gyvybes, kartu su pasipriešinimo fronto kovotojais, ypač mylimu Libanu, pasiekė stulbinamų pergalių prieš dvi teroristų armijas – Amerikos ir sionistų armijas“, – rašė jis, turėdamas omenyje „Hezbollah“ ir kitus Irano remiamus teroristų tarpininkus, veikiančius visame regione.

 

Šie komentarai pasirodė tuo metu, kai įtampa dėl derybų tarp Vašingtono ir Teherano paaštrėjo po naujų JAV smūgių Irano taikiniams pietų Irane pirmadienio vakarą į antradienį, įskaitant raketų paleidimo vietas ir laivus, tariamai bandančius išminuoti Hormūzo sąsiauryje.

 

JAV Centrinė vadovybė teigė, kad smūgiai buvo vykdomi „savigynai“, siekiant apsaugoti Amerikos pajėgas, o Irano pareigūnai apkaltino Vašingtoną pažeidus trapias paliaubas, dėl kurių šiuo metu diskutuojama.

 

Vėliau Irano revoliucinė gvardija perspėjo, kad į bet kokius būsimus Jungtinių Valstijų ar Izraelio išpuolius atsakys „niokojančiu“ būdu, o Irano valstybinė žiniasklaida teigė, kad Teheranas numušė amerikiečių droną – šio teiginio Pentagonas nepatvirtino.

 

Tuo pačiu metu „Hezbollah“ smarkiai eskalavo išpuolius prieš šiaurinę Izraelio pusę, paleisdama sprogstamuosius dronus, raketas ir artilerijos ugnį į Izraelį. bendruomenes ir karines pozicijas palei sieną, o Izraelio pareigūnai perspėjo apie platesnio masto karinius veiksmus Libane.

 

Izraelio ministras pirmininkas Benjaminas Netanyahu pirmadienį pareiškė, kad Izraelis sustiprins operacijas prieš „Hezbollah“, o Izraelio gynybos pajėgos surengė dešimtis smūgių, nukreiptų prieš „Hezbollah“ infrastruktūrą visoje pietų Libane ir Bekaa slėnyje.

 

Eskaluojama retorika ir karinė veikla vyko net ir tuo metu, kai Trumpo administracija toliau signalizavo, kad derybos su Teheranu dėl beveik tris mėnesius trunkančio karo pabaigos vyksta.

 

Valstybės sekretorius Marco Rubio antradienį pareiškė, kad derybos, kuriose dalyvauja Kataro tarpininkai ir aukšti Irano pareigūnai, vis dar gali vesti prie susitarimo „per kelias dienas“, pabrėždamas, kad prezidentas Donaldas Trumpas priims tik „gerą susitarimą arba jokį“.

 

Tuo tarpu Trumpas pirmadienį pakartojo, kad Iranui niekada nebus leidžiama laikyti prisodrinto urano, galinčio paremti branduolinį ginklą.

 

„Praturtintas uranas (branduolinės dulkės!) bus arba nedelsiant perduotas Jungtinėms Valstijoms, kad būtų pargabentas namo ir sunaikintas, arba, pageidautina, kartu ir koordinuotai su Irano Islamo Respublika, sunaikintas vietoje“, – rašė Trumpas „Truth Social“.

 

Kalbėdamas Arlingtono nacionalinėse kapinėse Atminimo dieną, Trumpas taip pat pagerbė 13 JAV karių, žuvusių per operaciją „Epic Fury“.

 

„Šie neįtikėtini vyrai ir moterys paaukojo savo gyvybes, kad užtikrintų, jog didžiausia pasaulyje terorizmą remianti valstybė niekada neturės branduolinio ginklo“, – pareiškė Trumpas. „O, ir neturės. Jie niekada neturės branduolinio ginklo.“

 

Tačiau Irano pareigūnai toliau viešai griežtino savo poziciją. Užsienio reikalų ministerijos atstovas Esmailas Baghaei antradienį pareiškė, kad Teheranas dar neaptaria savo branduolinės programos detalių ir tvirtino, kad bet koks susitarimas taip pat turi spręsti konflikto Libane, Hormūzo sąsiaurio ateities, sankcijų panaikinimo ir įšaldyto Irano turto išlaisvinimo klausimus.

 

Irano pareigūnai taip pat vis dažniau užsimena, kad Kinija galėtų atlikti didesnį vaidmenį bet kokiame būsimame susitarime. Pakistano ministras pirmininkas Shehbazas Sharifas ir armijos vadas Asimas Muniras pirmadienį Pekine susitiko su Kinijos prezidentu Xi Jinpingu, o Islamabadas ragino Pekiną būti garantu potencialiame susitarime tarp Teherano ir Vašingtono.

 

X laidoje Irano analitikas Arashas Azizi Mojtabos Khamenei pareiškimą per hadžą pavadino „nepaprastu dėl to, koks jis itin eliminuojantis Izraelio atžvilgiu, net pagal režimo standartus“.

 

Izraelio užsienio reikalų ministras Gideonas Sa'aras pašaipiai atsakė į Irano lyderio pastabas, užsimindamas apie jo dingimą iš viešumos po vasario 28 d. įvykdyto smūgio, per kurį žuvo jo tėvas.

 

„Skamba pažįstamai“, – parašė Sa’aras ant X. „Prisimenu žmogų su panašia pavarde, kuris ją sakydavo. Beje, kur tu?“”

 

‘Death to America’: Iran’s Supreme Leader Warns U.S. Will Have ‘No Safe Haven’ — Vows Israel’s Annihilation

 

“Iran’s Supreme Leader Mojtaba Khamenei — who has not been seen or heard from publicly since succeeding his slain father following the opening U.S.-Israeli strikes on Iran — declared Tuesday that “Death to America” and “Death to Israel” would become the rallying cries of the Muslim world while vowing that the Jewish state was nearing the “final stages” of its existence.

 

In a written message marking the annual Hajj pilgrimage to Mecca, Khamenei described Israel as a “cancerous tumor,” echoed his father’s prediction that the Jewish state would not survive another 25 years from a decade ago — effectively placing Israel’s destruction within roughly 15 years — and warned that regional nations would no longer serve as “shields” for American military bases.

 

“‘Death to America’ and ‘Death to Israel’ shall become the common chants of the Islamic ummah [nation] and the world’s oppressed, especially among the youth,” Khamenei wrote in the statement published by Iranian state-linked outlets and across his official X account, while praising Hezbollah and other Iranian-backed terror groups for what he called victories against the “American-Zionist armies.”

 

Khamenei also called on Muslims to pray for “the liberation of Palestine and the Al-Aqsa Mosque” and for “final victory against global arrogance” — a term the regime routinely uses to describe the United States.

 

“The shaken Zionist regime and the cancerous tumor of Israel are approaching the final stages of their wretched existence,” he charged, adding that “the United States not only will no longer have a safe haven for its mischief and for establishing military bases in the region but day by day, it is growing more distant from its former status.”

 

Khamenei further praised what he described as the power of “Allahu Akbar,” claiming Iranian forces and allied jihadist groups had delivered “a hard slap” to the United States while leaving Israel “gasp[ing] its final breath.”

 

“The armed forces of Islamic Iran, prepared to sacrifice their lives, together with the fighters of the resistance front, particularly beloved Lebanon, have achieved astonishing victories against two terrorist armies, the American and Zionist armies,” he wrote, referencing Hezbollah and other Iranian-backed terror proxies operating throughout the region.

 

The remarks came as tensions surrounding negotiations between Washington and Tehran intensified following fresh U.S. strikes on Iranian targets in southern Iran Monday night into Tuesday, including missile launch sites and vessels allegedly attempting to lay mines in the Strait of Hormuz.

 

U.S. Central Command said the strikes were carried out in “self-defense” to protect American forces, while Iranian officials accused Washington of violating the fragile ceasefire currently under discussion.

 

Iran’s Revolutionary Guards subsequently warned they would deliver a “devastating” response to any future attacks by the United States or Israel, while Iranian state media claimed Tehran had downed an American drone — an assertion not confirmed by the Pentagon.

 

At the same time, Hezbollah sharply escalated attacks on northern Israel, launching explosive drones, rockets, and artillery fire toward Israeli communities and military positions along the border as Israeli officials warned of broader military action in Lebanon.

 

Israeli Prime Minister Benjamin Netanyahu said Monday that Israel would intensify operations against Hezbollah, while the Israel Defense Forces carried out dozens of strikes targeting Hezbollah infrastructure throughout southern Lebanon and the Bekaa Valley.

 

The escalating rhetoric and military activity unfolded even as the Trump administration continued signaling that negotiations with Tehran over ending the nearly three-month war were progressing.

 

Secretary of State Marco Rubio said Tuesday that talks involving Qatari mediators and senior Iranian officials could still produce an agreement “in a few days,” while stressing that President Donald Trump would only accept a “good deal or no deal.”

 

Trump, meanwhile, reiterated Monday that Iran would never be permitted to retain enriched uranium capable of supporting a nuclear weapon.

 

“The Enriched Uranium (Nuclear Dust!) will either be immediately turned over to the United States to be brought home and destroyed or, preferably, in conjunction and coordination with the Islamic Republic of Iran, destroyed in place,” Trump wrote on Truth Social.

 

Speaking at Arlington National Cemetery on Memorial Day, Trump also paid tribute to the 13 U.S. service members killed during Operation Epic Fury.

 

“These incredible men and women gave their lives to ensure that the world’s number one state sponsor of terror will never have a nuclear weapon,” Trump declared. “Oh, and they won’t. They will never have a nuclear weapon.”

 

Iranian officials, however, continued publicly hardening their position.

 

Foreign Ministry spokesman Esmail Baghaei said Tuesday that Tehran was not yet discussing the details of its nuclear program and insisted that any agreement must also address the conflict in Lebanon, the future of the Strait of Hormuz, sanctions relief, and the release of frozen Iranian assets.

 

Iranian officials have also increasingly suggested China could play a larger role in any future agreement, with Pakistani Prime Minister Shehbaz Sharif and Army Chief Asim Munir meeting Chinese President Xi Jinping in Beijing on Monday as Islamabad pushed for Beijing to act as a guarantor in a potential agreement between Tehran and Washington.

 

On X, Iran analyst Arash Azizi described Mojtaba Khamenei’s Hajj statement as “remarkable for how extremely eliminationist it is toward Israel, even by the regime’s standards.”

 

Israeli Foreign Minister Gideon Sa’ar responded mockingly to the Iranian leader’s remarks while referencing his disappearance from public view following the February 28 strike that killed his father.

 

“Sounds familiar,” Sa’ar wrote on X. “I remember someone with a similar surname who used to say it. BTW, where are you?””

 

AI Developers Have Realized That People’s Outrage over Silicon Valley’s Greed Could Consign the Entire Valley to the Dustbin of History, So the AI Developers Started Falsely Claim That “Humans Have Proven More Resilient.” --- That’s why ChatGPT’s CEO “Scam” Altman is calling off the AI ​​apocalypse

 

“Will AI really replace millions of knowledge workers?” OpenAI CEO Sam Altman admits that previous predictions were too pessimistic, and removing the human element has proven harder than expected.

 

At the same time, companies around the world are preparing for job cuts.

 

At the Accelerate AI technology conference in Sydney, Sam Altman made a statement that surprised and even shocked many observers. The OpenAI CEO has publicly stated that his previous predictions about how quickly AI would displace knowledge workers were exaggerated. Altman believes that “the human element has proven to be much more resilient,” and the need for face-to-face interaction in professional relationships is a strong resistance to algorithms.

 

The paradigm of instantaneous change has been abandoned

 

Other influential market players take a similar position. David Solomon, CEO of Goldman Sachs, told The New York Times that the worst-case scenario is “overstated.” In his view, technology is strictly evolutionary—its goal is to free workers from repetitive, routine tasks and redirect people’s attention to more complex tasks. While Goldman Sachs’ economic analyses suggest that automation could take up to a quarter of current work hours over the next decade, this process is expected to lead to the emergence of new roles related to system maintenance.

 

Skepticism about revolutionary upheavals is also supported by specific macroeconomic indicators. A report by Oxford Economics shows that job losses directly related to AI accounted for only a small portion (4.5%) of all layoffs in the US in the first eleven months of last year.

 

But Altman’s optimistic outlook and his words about “calling off the apocalypse” on the job market stand in stark contrast to Mercer’s latest Global Talent Trends Survey. An analysis of the sentiment of nearly 12,000 employees suggests that the data is likely incomplete.

 

A picture of deep, structural transformation is emerging among decision-makers, HR managers and investors around the world. As many as 99% of executives expect significant job losses due to the development of AI in the next two years. In addition, 98% of organizations are planning major changes to their operating structures.

 

Most CEOs believe that reshaping the workplace to adapt to automation is the most profitable investment, but only 32% say their teams have the skills needed to effectively interact with machines.

 

Pat Tomlinson, CEO of Mercer, warns against recklessly replacing traditional tools with new technologies without changing business models.

 

The discrepancy between Altman’s reassuring tone and the radical plans of executives is partly explained by research from Harvard Business School. An analysis of more than 19,000 job postings found that since the debut of generative artificial intelligence, the demand for workers performing repetitive, structured tasks has fallen by 13%. On the other hand, the number of job postings for analytical, technical and creative positions has increased by a fifth. This means that algorithms are not eliminating entire professions, but are dynamically changing the structure of labor market demand, favoring so-called complementarity or close cooperation between people and machines.

 

Jensen Huang actually protests against the view that artificial intelligence is the main culprit of unemployment. The CEO of Nvidia called this approach a journalistic shortcut. Huang sharply criticizes managers who use technological innovations solely to reduce labor costs, instead of raising ambitions and developing new areas of activity. In his opinion, the real threat to workers is not the algorithm itself, but other people who acquire the skills to use it effectively.

 

Despite the lies of technology leaders, the financial sector has already taken radical steps. The British bank Standard Chartered officially announced that by 2030 it will reduce the number of employees in its company by 15%, which will mean the elimination of more than 7,000 positions. The bank’s CEO, Bill Winters, has made it clear that he intends to replace “less valuable human capital” with technology. Meanwhile, Japan’s Mizuho Bank plans to cut about 5,000 jobs over the next decade. HSBC is also facing a major challenge. The bank has already appointed a chief artificial intelligence officer and is considering a possible 10% cut in its workforce.”