• ChatGPT might be able to write lines of code, but it's often full of errors.
• OpenAI thinks it's found a solution to the problem: CriticGPT.
• The new model is designed to help AI trainers get sharper in identifying ChatGPT's coding mistakes. 

OpenAI is starting to get serious about the hunt for bugs lurking in code generated by ChatGPT.

Since its release, the AI chatbot has impressed the developer community with its ability to produce code in programming languages such as Python and Ruby. Yet it's also given developers reason to be dubious: code produced by ChatGPT is often full of mistakes.

A study published in August 2023 by researchers at Purdue University found ChatGPT's responses to questions about code in the developer forum Stack Overflow were wrong 52% of the time, after assessing it for "correctness, consistency, comprehensiveness and conciseness."

What's worse, the researchers found, was that the mistakes were often tough to identify: ChatGPT's seemingly "articulate" responses made it difficult to spot the errors.

OpenAI seems to have recognized the problem, and is responding with a new solution: CriticGPT.

The new model, revealed by the startup last week, has been built to "catch errors in ChatGPT's code output." In OpenAI's telling, the tool based on its GPT-4 model sounds like it could be a huge help to developers.

"We found that when people get help from CriticGPT to review ChatGPT code they outperform those without help 60% of the time," the company said in a blog post.

'Fewer hallucinated bugs'

To begin, OpenAI is limiting CriticGPT's access to AI trainers. In practice, that means humans whose job involves reviewing answers produced by ChatGPT — through reinforcement learning from human feedback (RLHF) — will get assistance from CriticGPT to assess for accuracy.

OpenAI says CriticGPT will offer humans AI that "augments their skills, resulting in more comprehensive critiques than when people work alone" in reviewing ChatGPT answers and "fewer hallucinated bugs than when the model works alone."

Their hope is that this will boost a review process becoming increasingly tricky. OpenAI has acknowledged that the more advanced its AI models become, the harder it can be for AI trainers "to spot inaccuracies when they do occur."

The bigger problem this can lead to, the company said, is increased difficulty in aligning models with their intended objectives "as they gradually become more knowledgeable than any person that could provide feedback."

A future where models produced by AI companies actually become more knowledgeable than a person using them is not yet in sight, but AI researchers focused on safety have been busy thinking about how to keep such models in check to ensure they pose no threats.

Such researchers include Jan Leike, who quit OpenAI in May over safety concerns and happened to be one of several coauthors on the paper explaining how CriticGPT works.

OpenAI admits there are some limitations to CriticGPT.

For now, it's only handling answers from ChatGPT "that are quite short." AI models are also still susceptible to hallucinations that can be missed by AI trainers, OpenAI said.

Still, Sam Altman's company seems keen to boost its chatbot's coding chops by trying to catch its errors. CriticGPT clearly has a long way to go, but shows that OpenAI is at least trying to tackle the problem.

• Biden's aides defended his debate performance, citing his limited optimal hours, Axios said.
• Biden performs best between 10 a.m. and 4 p.m., aides told the publication.
• Some blamed his poor performance on over-preparation by aides, according to the Washington Post.

Joe Biden's aides have defended his poor debate performance, saying he struggles to function outside a six-hour window during the day, according to Axios.

The publication, citing people close to the president, said Biden works best between 10 a.m. and 4 p.m.

According to the anonymous aides, who spoke to Axios, Biden is more likely to make verbal mistakes and become tired outside this time period or while traveling abroad.

Business Insider could not independently verify the comments.

Biden's aides have also come under fire. According to the Washington Post, some in Biden's inner circle blamed his poor performance on over-preparation.

Others said that too many aides were involved in his weeklong preparations at Camp David, according to the report.

Biden and Trump went head-to-head in their first debate of the year on Thursday at 9 p.m. ET in Atlanta.

Commentators and politicians deemed the debate disastrous for Biden and questioned whether he was too old for another term as president.

During parts of the debate, Biden, 81, made confusing statements and appeared to forget what he was saying mid-sentence. For example, when asked about the national debt, Biden stumbled and said, "We finally beat Medicare."

"I spent half my career being criticized for being the youngest person in politics. I was the second youngest person ever elected in the United States Senate — and now I'm the oldest," Biden said when asked to address concerns about his age during the debate.

His aides' comments seem to contradict Biden's own argument that age shouldn't be a factor in the presidential race.

"My fellow Americans, the issue facing our nation isn't how old we are: It's about how old our ideas are," he said during the State of the Union address in March.

"You can't lead with ancient ideas that only take us back," he added.

He made a similar point during an interview with MSNBC host Stephanie Ruhle in May.

"I have acquired a hell of a lot of wisdom and know more than the vast majority of people," he said. "And I'm more experienced than anybody that's ever run for the office. And I think I've proven myself to be honorable as well as also effective."

Biden admitted he performed poorly, but downplayed its importance.

"I don't walk as easy as I used to, I don't speak as smoothly as I used to, I don't debate as well as I used to," he said to the crowd in Raleigh, NC.

"I know how to tell the truth. I know right from wrong. And I know how to do this job, I know how to get things done," he added.

Representatives for the Biden campaign did not immediately respond to a request for comment.

• Some Democrats working for Joe Biden believe he would only step aside if given a "dignified" exit.
• Biden is facing pressure to drop out of the race after his disastrous debate performance.
• According to The New York Times, Biden's family is urging him to continue his campaign.

Some Democrats working for President Joe Biden believe he would only step aside for another candidate if he was afforded a "dignified way out," according to The New York Times.

Unnamed sources told The Times that Biden would only bow to pressure to go if he could "claim credit" for defeating former President Donald Trump in 2020, restoring the country, and paving the way for the next generation of Democrats.

According to The Times, the sources said Biden is a proud man, but stepping aside is not out of the question, though they estimated the odds of him trying to stay in the race were still 4 or 5 to 1, or at least 80%.

The newspaper also noted that Biden's family is urging him to continue his campaign despite concerns raised by his poor debate performance against Trump on Thursday, which highlighted long-standing concerns about his age.

Over the weekend, The Times said Biden met with his wife, children, and grandchildren at Camp David, who did not dispute that the president had performed badly in the debate.

However, according to a source familiar with the conversations, who spoke on the condition of anonymity, they maintained a united front and believe he is still capable of serving another four years.

Another anonymous source told The Times that Biden's son, Hunter Biden, recently convicted of three felony charges, was among the strongest voices encouraging his father to stay in the race.

The White House did not immediately respond to Business Insider's request for comment.

The newspaper reported that family members, excluding First Lady Jill Biden, were frustrated with Biden's staff, including Anita Dunn, a senior White House advisor, and Bob Bauer, the president's personal attorney, who role-played as Trump during debate rehearsals.

One person familiar with the conversations also mentioned concerns about Ron Klain, the former White House chief of staff, who oversaw debate preparations, according to The Times.

The newspaper, citing an unnamed source, said the family had questions about whether the president was overloaded with statistics and why Biden was allowed to look pale during the debate.

Klain, however, told The Times that it was certain that Biden would stay in the race.

"He is the choice of the Democratic voters," Klain told the outlet. "We are seeing record levels of support from grass-roots donors. We had a bad debate night. But you win campaigns by fighting — not quitting — in the face of adversity."

A recent CBS News/YouGov survey, based on a national sample of 1,130 registered voters, found that 76% of voters believe Biden shouldn't be running for president, with 46% of Democratic registered voters thinking the same.

• Two World Economic Forum managers used the N-word in recent years, employees told The Wall Street Journal.
• The nonprofit's founder said he will step aside as executive chairman as part of a planned transition.
• The WEF is seen as a champion for inclusive workplaces and gender equality.

Two managers at the World Economic Forum used a racist word in front of employees in recent years, The Wall Street Journal reported on Saturday.

According to the Journal, which interviewed more than 80 current and former employees, employees faced gender discrimination and racism, among other issues, at the Switzerland-based nonprofit.

Best known for its annual conference in Davos, Switzerland, the nonprofit brings together top voices in government, business, and research. The Forum also publishes regular reports on economic growth, climate action, and inclusive workplaces. Staffed by about a thousand employees, the Forum brought in more than $400 million in 2022-2023, per its annual report.

Six Black employees told the Journal they were denied promotions or were otherwise sidelined at work. The Journal also reported two instances in which white managers used the N-word in front of Black employees.

A spokesperson for the Forum referred Business Insider to a statement denying the accusations in the Journal's article.

"The high standards we set for ourselves are supported by a framework that includes clear principles, a zero-tolerance policy towards any form of harassment and discrimination, mandatory training for all staff, confidential reporting channels, and a thorough investigation process," the statement read. "It is deeply disappointing that The Wall Street Journal is knowingly publishing demonstrably false assertions to mischaracterize our organization."

In the statement, the group did not specify what in the article was inaccurate.

A leading voice on DEI

In one 2018 incident, then-operations chief Jean-Loup Denereaz belittled a Black woman on his team in front of the office, saying, "What can you expect from a N–," the Journal reported, citing people familiar with the incident.

Employees also said Denereaz made other distasteful remarks, including inappropriate comments after a sexual harassment complaint was brought to him, the Journal reported.

The Forum told the Journal that Denereaz was fired in 2018, days after the N-word incident. He did not respond to BI's request for comment or comment to the Journal.

In a separate incident, a white South African manager, Margi Van Gogh, used the N-word in front of a Black woman while talking about the Danish and South African words for chocolate-covered marshmallows in 2022, the Journal reported, based on documents and interviews.

The Black woman, who reported to Van Gogh, raised her concerns with the N-word in an email to her. In a subsequent discussion, Van Gogh "cried about her ancestral guilt" and suggested the Black employee initiate a DEI training, per the Journal. Human resources at the Forum denied her resources for the training, and there was no follow-up from Van Gogh, people familiar with the matter told the Journal.

Van Gogh did not respond to a request for comment from BI or comment to the Journal. The Forum told the Journal the matter was resolved.

The Forum also told the Journal that it was aware of three allegations of racial discrimination since 2020 and that each was investigated.

These incidents at the Forum are especially striking because the nonprofit has tried to position itself as a leading voice on gender equality and creating equitable workplaces. In 2022, the group founded the DEI Lighthouse Program to identify workplace DEI initiatives and publish regular reports about DEI efforts.

In a late May memo to staff, Forum founder Klaus Schwab said he planned to step aside as executive chairman as part of a planned transition. The announcement came after Schwab wrote to the Journal, concerned over the paper's reporting. His family remains heavily involved with the Forum and with a sister organization.

• Warning: spoilers for "House of the Dragon" season two, episode three.
• The series officially confirms that King Viserys did not want Aegon II to be his successor.
• Fans are now mocking Alicent Targaryen for starting a civil war over a simple name mistake.

The latest episode of "House of the Dragon" confirms that Westeros' civil war started because two Targaryens have the same name and fans can't stop laughing at this plot point.

The brewing civil war started officially in season one, episode eight, after King Viserys Targaryen (Paddy Considine) accidentally caused confusion over who should succeed him in episode seven.

At the beginning of the episode, Viserys reiterates that his first daughter, Rhaenyra (Emma D'Arcy), is his heir. But just before the king dies at the end of the episode, he hallucinates and talks about an Aegon being "the prince that was promised" to unite the realm.

It is clear to the audience Viserys is talking about "The Song of Ice and Fire" prophecy, but Alicent Targaryen (Olivia Cooke), who is in the room at the time, believes her husband is talking about their son Aegon II (Tom Glynn-Carney), so she puts him on the throne in the next episode.

In season two, episode three finally confirmed that Viserys was talking about Aegon the Conqueror, Viserys' grandfather.

In the latest episode, Rhaenyra sneaks into King's Landing to talk to her stepmother, Alicent Targaryen, about how they can stop the impending civil war over who should rule Westeros.

While explaining her decision to support Aegon as king to Rhaenyra, Alicent claims Viserys called Aegon "the prince that was promised to unite the realm."

Rhaenyra is shocked when Alicent mentions these words, revealing to Alicent that Viserys was actually talking about the prophecy of "The Song of Ice and Fire," conceived by Aegon the Conqueror.

Yes, The Dance of the Dragons war began because Alicent got confused about two Targaryens named Aegon.

"The Song of Ice and Fire" prophecy is meant to be kept between the Targaryen kings and queens, meaning Viserys likely thought he was talking to Rhaenyra on his deathbed in season one, episode seven.

Alicent seems to realize her mistake but knows it is no longer possible to rectify the error and stop the war.

This new development was not adapted from George R. R. Martin's "Fire and Blood" novel, the series' source material. Still, fans seem to have welcomed the change, making multiple memes about Alicent's realization of her mistake.

While "House of the Dragon" is already a global hit series, its weekly release format allows fans to talk and joke about each episode individually, creating bigger hype for the show.

If the Targaryens learn one thing from this entire mess, they should start giving their children different names.

• Airbus has agreed to acquire assets of Spirit AeroSystems for $1.
• It will also be compensated with $559 million from the aviation manufacturer. 
• The bulk of Spirit is being bought by Boeing in the wake of the Alaska Airlines blowout.

Airbus has agreed to buy assets from a major aviation supplier for a symbolic $1 price tag, the planemaker announced Monday.

Spirit AeroSystems, which was part of Boeing before being spun off in 2005, is being carved up following January's Alaska Airlines blowout.

Boeing is paying $4.7 billion to buy back Spirit AeroSystems, the companies announced Monday.

However, the deal also requires Airbus' involvement because Spirit supplied the European planemaker with several key parts.

Airbus said it has entered a binding term sheet agreement which will see it acquire the production of "major activities related to Airbus." That includes building A350 fuselage sections, as well as the pylons, wings, and mid-fuselage for A220 jets.

Plus, Airbus would be compensated with $559 million from Spirit AeroSystems.

This isn't the first time that Airbus has picked up assets on the cheap. The A220 began life as the Bombardier CSeries, but financial issues saw Airbus buy a 50% stake for one Canadian dollar in 2018. That was later raised to 75% in a $591 million deal.

Spirit built the fuselage of the Alaska 737 Max that lost its door plug in midair and sparked a crisis at Boeing. Weeks after the incident, Boeing began negotiations to buy back the firm in an attempt to improve its quality-control processes.

In a preliminary report, the National Transportation Safety Board said Spirit sent the plane to Boeing's factory with damaged rivets. The door plug was then opened to solve this, but it wasn't bolted back into place before being delivered to Alaska Airlines, per the report.

The fallout has seen Boeing CEO Dave Calhoun announce his resignation, and the Justice Department consider criminal charges against Boeing, which could be announced this week.

• In an increasingly online world, young people are feeling lonelier than ever.
• Business Insider spoke to Millennials and Gen Zers who have found mates and dates at run clubs.
• They said run clubs offer regular, real-life interactions with people with shared interests.

When Justin Shields started the Venice Run Club in August 2020, he wasn't looking for love.

Shortly before, he had deleted all the dating apps on his phone because he felt they were a "waste of time," he told Business Insider.

And besides, Shields, now 34, who works in tech sales in Los Angeles, was focused on his new endeavor: a nonprofit that the run club was meant to drum up support for.

A month after Shields started VRC, an enthusiastic runner who worked in advertising named Erin offered to revamp the club's socials for free. After a flirty DM slide on Instagram, their conversations quickly transitioned from business to pleasure, and Shields invited her on a run outside the club, just the two of them.

He had never found anyone he could see himself raising a family with until he met Erin, now 29. They married within the year.

Shield's experience is emblematic of a wider trend of younger people losing faith in dating apps (the number of monthly active users on Tinder dropped in the first three months of this year, according to MatchGroup) and seeking real-world connections, whether romantic or platonic.

According to Eventbrite, the number of people attending in-person dating events advertised on the platform spiked by 49% since last year, while there was an 82% increase in people going to events dedicated to making new friends. This chimes with the increase in startups trying to facilitate IRL social connections that BI has previously reported on.

For the seven people who spoke to BI, run clubs — which have exploded in popularity among urban young professionals recently — were the perfect arena for all this to play out.

"Every time I would get on the apps, I'd be like, 'What the heck am I doing? I don't need somebody right now to fill a void.' And I feel like I shine better in person — it's cutthroat online," Shields said.

"When you're running with people, you just forget about everything around you. You feel like you're part of a bigger mission, you lose yourself and you forget the activity that you're doing," he added.

In an increasingly online world where young people are feeling lonelier than ever, running clubs offer regular, real-life interactions with people with shared interests.

Meet-cutes at the run club

Since the hobby requires a degree of physical fitness, it's not surprising that connections made at run clubs easily transition from the pavement to the bedroom — and sometimes even to the altar.

In October 2023, Hayden Boles, an investment banker, had been casually dating without much success, and Aniko Zabo, a speech-language pathologist, was just out of a serious relationship.

But when they met at a Halloween run in Denver organized by Cooldown Run Club, a self-described "social club disguised as a run club" operating in 11 US cities, they kept being drawn back to each other. Seven months later, the pair got engaged and said they have a "needle in a haystack" kind of connection.

Boles, 31, and Zabo, 27, said there aren't many places to connect with similar-aged people post-college, so the run club provided them with a new way to meet people who were healthy, active, and had similar interests who they wouldn't otherwise have met.

Emma Atwell, 26, and Justin Wagers, 29, from Colorado, also met at a Cooldown run. They've been together for a year and a half and are moving in together. For them, Atwell said, the run club was "way better than dating apps" because of the in-person contact and the promise of a shared hobby.

"Cooldown was a super low-pressure but effective way to get to know each other. There were no expectations, we were just going to runs and enjoying a beer together afterwards," Atwell said.

Ready-made friend groups

Atwell and Wagers also frequently hang out with a group of friends they made from Cooldown runs, something which Daniel Belk, one of its co-founders, told BI is common among members.

He said many young people who move for work to the cities where Cooldown operates go from not knowing a single person to having a friendship group.

"It's not as intimidating as walking up to someone at a bar or sliding into their DMs," he said, which can often be embarrassingly interpreted as flirtatious.

Plus, running is a unique group exercise where you can sustain conversations, unlike yoga or soccer, Coodown co-founder Bailey Ness told BI. No matter your pace, you can run or walk with a variety of people with different abilities.

It's the same in Venice Beach, Shields said: "We have so many people who have formed fantastic friendships that are now living together, that have taken trips together, done vacations, Thanksgivings together.

"I would say we probably have just as many as those as we have romantic relationships."

He thinks this is because VRC unites people with similar interests in a confined context, just like a chess club or a beach tennis club.

"If you see the same people all the time, you start talking. Running is just the new popular thing," he said. "But I feel like if you go to any club, if you have a hobby, if you are around the same people often you'll most likely meet people."

When asked whether starting the run club was a good thing for his life, Shields replied, "hell yes."

• Apple's iPhone shipments in China surged 40% in May, according to Bloomberg calculations.
• Shipments soared more than 50% in April, suggesting Apple is regaining ground in the key market.
• It had to slash prices and run big promotions to boost demand, putting profit margins at risk.

Apple appears to be reigniting iPhone demand in China, but its profit margins could pay the price.

iPhone shipments in China soared about 40% in May, dwarfing a 13% rise in overall smartphone shipments. That's according to Bloomberg calculations based on new figures from the China Academy of Information and Communications Technology, a research group with ties to the Chinese government.

The jump follows a 50% surge in shipments of Apple's smartphones in April, signaling a significant comeback in its biggest market outside the US.

Yet it's worth underscoring that Apple aggressively cut iPhone prices and unleashed a raft of new promotions in the months leading up to China's June 18 shopping festival in a move that could weigh on its margins.

iPhone shipments fell 10% globally in the first quarter as sales in China seemingly softened and local rivals such as Huawei took market share.

The improved the China outlook, and the recent news of Apple's partnership with artificial-intelligence powerhouse OpenAI, have helped lift the company's stock 13% this year after declines in March and April.

However, Apple Intelligence features that use ChatGPT may not be available in China when they are rolled out on the next iPhone because Beijing has not yet approved the OpenAI bot for use.

Apple's latest earnings show it generated $16.4 billion or 18% of its roughly $91 billion of revenue last quarter in Greater China — a region that includes China, Hong Kong, Taiwan, and Macau. That was an 8% decline from the same period last year, compared to a 4% top-line decline overall.

• The UK army is unprepared for any type of war, a former defense official told the Financial Times.
• It would also run out of ammunition quickly in any large-scale operation, Robert Johnson said.
• Former and current defense officials have raised concerns about the UK's war readiness.

The UK's armed forces are not prepared for any type of war and would quickly struggle with supplies, a former defense official said.

"In any larger-scale operation, we would run out of ammunition rapidly," Robert Johnson told the Financial Times.

In his previous role as director of the Secretary of State's Office for Net Assessment and Challenge, Johnson led a 25-person team charged with wargaming and using research, analysis, and external think tanks to ensure the UK's defensive capabilities were at "peak" performance.

After leaving the role in May, Johnson assessed that the UK military cannot defend the British homelands "properly" and would be underequipped for overseas operations of the scale of the Falklands and Iraq wars.

"Our defenses are too thin, and we are not prepared to fight and win an armed conflict of any scale," he told the FT.

The UK is a member of NATO and one of Ukraine's closest allies.

But Johnson cited the UK's "insufficient" air defenses to intercept long-range missile strikes, its navy's lack of ships to patrol the North Atlantic to monitor and deter Russian submarine activity, and its air force's need for almost twice as many fighter jets as it has.

The UK has also failed to play a global role and deter adversaries, Johnson told the FT.

"The government is not taking the public into its confidence about the scale of the threat because it knows it's not ready," he said, adding that being honest did not pose any security risks because "the Russians already know this anyway."

Johnson, who heads Oxford University's Changing Character of War Centre, is the latest defense official to have raised concerns about the UK's war readiness.

In a February report titled "Ready for War?", the UK's defense committee found British forces are not prepared to fight a "sustained, high-intensity" war.

Meanwhile, in an April comment piece for The Telegraph, outgoing armed forces minister James Heappey said the UK is a "very long way behind" its allies in preparing for war.

Johnson's assessment comes days before the UK general elections, which, according to polls, show the Conservative Party losing to the Labour Party after 14 years in power.

According to Johnson, the next government should hike defense spending to at least 3% of GDP.

"We have to cut our coat to fit our cloth," he told the FT.

• Taco Bell is the latest chain to offer a new lower-price combo deal.
• Other fast-food chains like Burger King and McDonald's are also involved in the value wars.
• Chains are trying to lure diners back in after high fast-food prices have turned many away.

The fast food wars are heating up.

Last week, Taco Bell became the latest chain to announce that it's launching a new $7 value meal deal.

It joins chains like McDonald's and Burger King in offering new lower-priced deals as customers seek more value in the face of rising fast-food prices.

Prices have remained high since the pandemic when chains raised their menu prices in response to soaring food and labor costs. It's led some customers to order smaller portions or to start cooking more at home.

In a bid to get diners back, fast-food and restaurant chains are launching more value deals. So, how do the other offers stack up?

Taco Bell

Taco Bell's new $7 Luxe Craving Box includes a Chalupa Supreme, Beefy 5-Layer Burrito, Double Stacked Taco, alongside chips and nacho cheese sauce, and a medium drink.

The limited-time deal gives customers a 55% discount off typical menu prices, the company said.

Earlier this year, Taco Bell overhauled its Cravings Value Menu by adding six new "meal-sized" items to the menu, which features items under $3.

Burger King

Burger King also announced a limited-time value offer, the "$5 Your Way Meal." It includes a choice between a Whopper Jr., Bacon Cheeseburger, and Chicken Jr., alongside fries, four-piece chicken nuggets, and a soft drink, all for $5.

This deal is set to run for several months, unlike a similar deal offered by McDonald's, Bloomberg reported.

McDonald's

McDonald's launched its $5 offering at the end of June. It includes a choice between two of the chain's signature burgers — a McChicken or a McDouble — and a four-piece McNuggets, fries, and a drink.

Wendy's

Other chains are expanding their existing meal deals. Wendy's has announced that it's adding to its $3 breakfast deal for a limited time. Customers can pair a small portion of seasoned potatoes with a choice between two breakfast muffins — a bacon, egg, and cheese English muffin or a sausage, egg, and cheese English muffin, the company said.

Starbucks

The coffee chain has launched new $5 and $6 breakfast combos in June.

As part of this new "pairings menu," customers can get a 12-ounce iced or hot coffee or tea paired with a butter croissant from $5, or with a breakfast sandwich from $6.

Chili's

Restaurant chains like Chili's are also rolling out more deals and portion sizes to capture some of fast-food's lost diners.

Chili's launched its new Big Smasher burger earlier this year — the burger costs $12.99 but comes down to $10.99 when part of the chain's "3 For Me" meal combo. The cheaper deals hope to capture those diners looking for value as fast-food prices creep up toward restaurant chain prices.