Connect with us

Technology

Tony Fadell on mission-driven a**holes, Silicon Valley entitlement and why LLMs are ‘know-it-alls’

Published

on

Tony Fadell on mission-driven a**holes, Silicon Valley entitlement and why LLMs are 'know-it-alls'

Tony Fadell, the father of the iPod and founder of Nest, took the stage at TechCrunch Disrupt 2024 on Tuesday to talk about how building the next generation of deep tech startups requires mission-driven a**holes. The entrepreneur and investor did not hold back on stage as he called out Silicon Valley for its entitlement and dunked on LLMs being “know-it-alls”, earning a wave of laughs and applause across the fully-packed auditorium.

Fadell explained why he believes “mission-driven a**holes are a good thing, and in fact, needed to create and ship world class technology products.

“People work with people who are very difficult, and those are the ones that create and change the world. But there are two types of a**holes. Everybody’s an a**hole, but you gotta understand why,” Fadell said. “If they’re an a**hole, because it’s their ego, they’re trying to push people down, that is an egocentric a**hole. But, if you are an a**hole on the details, you’re sitting there pushing on the details, you’re not criticizing the people, but you are critiquing their work and saying you can do better, that is a mission-driven a**hole.”

Fadell thinks it’s not a bad thing when someone is keen on the details and makes sure their team is getting things right. He believes that focusing on details is what you need to make great products, and when you have a manager who cares, that’s a good thing.

Advertisement

The entrepreneur and investor also called out Silicon Valley for its entitlement, making a joke about how startups aren’t hiring Googlers because “you’re lucky that they even showed up.”

“They just showed up on a bus, and they came in for lunch, and then took the bus home,” Fadell said. “They say ‘I worked today, I’m getting a massage, oh, where’s the yogurt?’ That’s why I don’t like any of our startups hiring most Googlers because they’ve got this culture thing.”

He noted that during his days at General Magic in the 90s, the team decided they weren’t going to hire people from the East Coast because of their demands.

“We said, we will never hire people from the East Coast,” Fadell said. “So this was IBM and Sperry and all that s*** because they had to have their driver, or they had to have their company car, and they had to have their corporate lunch and their special executive toilet. We’re like we’re never hiring any of these people, it’s just not going to work, it’s a culture clash. And now I wake up today and Silicon Valley has turned into that s***, and I’m like get me the f*** out of here, yeah? Entitlement everywhere!”

Advertisement

Fadell then went on to call out LLMs for being know-it-alls, joking that no one wants to hire a know-it-all. He believes LLMs are great for certain things, but that they can’t be adopted across the board.

He believes LLMs can be great for entertainment, like if you’re asking ChatGPT to write you a funny poem. However, Fadell says that since LLMs are prone to hallucinations, they shouldn’t be used in cases where people could be at risk. For instance, he thinks doctors shouldn’t be using ChatGPT to create patient reports because people could be harmed if treatment plans are entered incorrectly or drugs are misnamed.

“If you look at artificial-specific models, they work really well,” Fadell said. “They don’t hallucinate, but LLMs are trying to be this general thing because we’re trying to make science fiction happen.”

Source link

Advertisement
Continue Reading
Advertisement
Click to comment

You must be logged in to post a comment Login

Leave a Reply

Technology

More than a quarter of new code at Google is generated by AI

Published

on

More than a quarter of new code at Google is generated by AI

Google is building a bunch of AI products, and it’s using AI quite a bit as part of building those products, too. “More than a quarter of all new code at Google is generated by AI, then reviewed and accepted by engineers,” CEO Sundar Pichai said on the company’s third quarter 2024 earnings call. It’s a big milestone that marks just how important AI is to the company.

AI is helping Google make money as well. Alphabet reported $88.3 billion in revenue for the quarter, with Google Services (which includes Search) revenue of $76.5 billion, up 13 percent year-over-year, and Google Cloud (which includes its AI infrastructure products for other companies) revenue of $11.4 billion, up 35 percent year-over-year.

Operating incomes were also strong. Google Services hit $30.9 billion, up from $23.9 billion last year, and Google Cloud hit $1.95 billion, significantly up from last year’s $270 million.

“In Search, our new AI features are expanding what people can search for and how they search for it,” CEO Sundar Pichai says in a statement. “In Cloud, our AI solutions are helping drive deeper product adoption with existing customers, attract new customers and win larger deals. And YouTube’s total ads and subscription revenues surpassed $50 billion over the past four quarters for the first time.”

Advertisement

Google is facing a potentially tough road ahead, however, following the August ruling that the company is a monopolist in the search and advertising markets. That case, brought by the US Department of Justice, is now in its remedies phase, and while there’s still a ways to go for the dust to settle, a Google breakup is on the table.

Source link

Continue Reading

Technology

Canon gives its mirrorless cameras a major video boost with trio of new pro hybrid RF lenses

Published

on

Canon RF 50mm F1.4L VCM and Canon RF 24mm F1.4L VCM lenses alongside the 35mm version on a rustic worktop

Canon has announced four new lenses for its mirrorless cameras, three of which are touted ‘hybrid’, meaning that they are designed for video and still creators. They are the RF 24mm F1.4L VCM, RF 50mm F1.4L VCM and the RF 70-200mm F2.8L IS USM Z.

The two pro-grade primes and zoom are made for Canon’s full-frame cameras and join two recently announced lenses of the same kind, the RF 35mm F1.4L VCM and RF 24-105mm F2.8L USM Z – the latter being quietly launched alongside the RF 200-800mm monster wildlife lens – bringing Canon’s range of hybrid lenses to five in all.

Source link

Advertisement
Continue Reading

Technology

How X users earn thousands from US election misinformation and AI images

Published

on

How X users earn thousands from US election misinformation and AI images
BBC A composite graphic showing an AI-generated image of Donald Trump as a military dictator with medals on his chest and a doctored image of Kamala Harris in a McDonald's uniform with a label saying fake and the X logo in the backgroundBBC

Some users on X who spend their days sharing content that includes election misinformation, AI-generated images and unfounded conspiracy theories say they are being paid “thousands of dollars” by the social media site.

The BBC identified networks of dozens of accounts that re-share each other’s content multiple times a day – including a mix of true, unfounded, false and faked material – to boost their reach, and therefore, revenue on the site.

Several say earnings from their own and other accounts range from a couple of hundred to thousands of dollars.

They also say they coordinate sharing each other’s posts on forums and group chats. “It’s a way of trying to help each other out,” one user said.

Some of these networks support Donald Trump, others Kamala Harris, and some are independent. Several of these profiles – which say they are not connected to official campaigns – have been contacted by US politicians, including congressional candidates, looking for supportive posts.

Advertisement

On 9 October, X changed its rules so the payments made to eligible accounts with a significant reach are calculated according to the amount of engagement from premium users – likes, shares and comments – rather than the number of ads under their posts.

Many social media sites allow users to make money from their posts or to share sponsored content. But they often have rules which allow them to de-monetise or suspend profiles that post misinformation. X does not have guidelines on misinformation in the same way.

While X has a smaller user base than some sites, it has a significant impact on political discourse. It raises questions about whether X is incentivising users to post provocative claims, whether they’re true or not, at a highly sensitive moment for US politics.

The BBC compared the approximate earnings reported by some of these X users with the amount they would be expected to earn, based on their number of views, followers and interactions with other profiles, and found them to be credible.

Advertisement

Among the misleading posts shared by some of these networks of profiles were claims about election fraud which had been rebutted by authorities, and extreme, unfounded allegations of paedophilia and sexual abuse against the presidential and vice-presidential candidates.

Some misleading and false posts that originated on X have also spilled on to other social media sites with a bigger audience, such as Facebook and TikTok.

In one example, an X user with a small following says he created a doctored image purporting to show Kamala Harris working at McDonald’s as a young woman. Other users then pushed evidence-free claims that the Democratic Party was manipulating images of its candidate.

Unfounded conspiracy theories from X about the July assassination attempt on Donald Trump were also picked up on other social media sites.

Advertisement

X did not respond to questions about whether the site is incentivising users to post like this, nor to requests to interview owner Elon Musk.

‘It’s become a lot easier to make money’

Freedom Uncut’s content creation lair – where he streams and makes videos – is decorated with fairy lights in the shape of an American flag. He says he is an independent, but would rather Donald Trump becomes president than Kamala Harris.

Free – as his friends call him – says he can spend up to 16 hours a day in his lair posting on X, interacting with the network of dozens of content creators he’s a part of, and sharing AI-generated pictures. He does not share his full name or real identity because he says his family’s personal information has been exposed online, leading to threats.

Advertisement

He is by no means one of the most extreme posters, and has agreed to meet me and explain how these networks on X operate.

He says he has had 11 million views over the past few months since he began posting regularly about the US election. He brings several up on the screen as we chat at his home in Tampa, Florida.

Free, a man with a full beard and closely cropped brown hair with a quiff on top, wearing a T-shirt with a stylised eagle and the slogan "America only". He is standing on a dock with boats behind him on the water

The man behind the Freedom Uncut account says he can earn in the “low thousands” from X

Some are obviously satire – Donald Trump looking like a character in The Matrix as he brushes aside bullets, or President Joe Biden as a dictator. Other AI images are less fantastical – including an image of someone on the roof of their flooded home as fighter jets pass by, with the comment: “Remember that politicians don’t care about you on November 5th.”

The image echoes Mr Trump’s claim that there were “no helicopters, no rescue” for people in North Carolina following Hurricane Helene. The claim has been rebutted by the North Carolina National Guard, which says it rescued hundreds of people in 146 flight missions.

Advertisement

Freedom Uncut says he sees his images as “art” that sparks a conversation. He says he is “not trying to fool anybody” but that he can “do so much more by using AI”.

Since his profile was monetised, he says he can make in the “low thousands” monthly from X: “I think it’s become a lot easier for people to make money.”

He adds that some users he knows have been making more than five figures and claims he could corroborate this by seeing the reach of their posts: “It’s at that point it really does become a job.”

He says it is the “controversial” stuff that tends to get the most views – and compares this to “sensationalist” traditional media.

Advertisement
A graphic showing two phones displaying screenshots of X posts from Freedom Uncut. One shows Donald Trump as Neo in the Matrix with the slogan "I can dodge bullets", while another shows a person on the roof of their flooded house as fighter jets fly overhead. A third shows President Joe Biden in a wheelchair with the US Capitol behind him and Nazi-like imagery surrounding them.

Freedom Uncut posts AI-generated images, which are often satirical, in support of Donald Trump or criticising Democrats

While he posts “provocative stuff”, he says it is “usually based in some version of reality”. But he suggests that other profiles he sees are happy to share posts they know not to be true. This, he says, is an easy “money-maker”.

Freedom Uncut dismisses concerns about false claims influencing the election, claiming the government “spreads more misinformation than the rest of the internet combined”.

He also says it is “very common” for local politicians to reach out to accounts like his on X for support. He says some of them have chatted to him about appearing on his live streams and spoken to him about creating and sharing memes, AI images and artwork for them.

Could any of these posts – misleading or not – have a tangible impact this election?

Advertisement

“I think that you’re seeing that currently. I think that a lot of the Trump support comes from that,” he says.

In Freedom Uncut’s view, there is “more trust in independent media” – including accounts sharing AI-generated images and misinformation – than in “some traditional media companies”.

‘No way to get to the truth’

Going head-to-head with the pro-Trump accounts Freedom Uncut describes are profiles such as Brown Eyed Susan, who has more than 200,000 followers on X.

Advertisement

She is part of a network of “die-hard” accounts posting content multiple times every hour in support of Democratic candidate Kamala Harris. While she uses her first name, she does not share her surname because of threats and abuse she has received online.

Speaking to me from Los Angeles, Susan says she never intended to start making money from her posts – or for her account’s reach to “explode”. Sometimes she posts and re-shares more than 100 messages a day – and her individual posts sometimes reach more than two million users each.

She says she only makes money from her posts because she was awarded a blue tick, which marks paid users on the site and some prominent accounts. “I didn’t ask for it. I can’t hide it, and I can’t return it. So I clicked on monetise,” she tells me, estimating she can make a couple of hundred dollars a month.

Susan, a woman with long blonde hair and half-rimmed glasses, sitting on a leather sofa with some abstract art behind her, on a video call with the BBC

Susan told the BBC in a video call the network of accounts she interacts with amplify each other’s posts to help Kamala Harris to win the election

Aside from posting about policy, some of her most viral posts – racking up more than three million views – have promoted unfounded and false conspiracy theories suggesting the July assassination attempt was staged by Donald Trump.

Advertisement

She acknowledges that a member of the crowd and the shooter were killed, but says she has genuine questions about Donald Trump’s injury, the security failings, and whether the incident has been properly investigated.

“There’s no way to get to the truth in this. And if they want to call it conspiratorial, they can,” she says.

Susan also shares memes, some of which use AI, taking aim at the Republican contender. Several more convincing examples make him look older or unwell. She says these “illustrate his current condition”.

Others show him looking like a dictator. She maintains that all her images are “obvious” fakes.

Advertisement

Like Freedom Uncut, she says politicians, including congressional candidates, have contacted her for support, and she says she tries to “spread as much awareness” as she can for them.

‘They want it to be real’

Following a row over whether Kamala Harris once worked at McDonald’s, a doctored image of her in the fast food chain’s uniform was shared on Facebook by her supporters and went viral.

When some pro-Trump accounts realised it was an edited photo of a different woman in the uniform, it triggered unfounded accusations that the image came from the Democratic Party itself.

Advertisement

An account called “The Infinite Dude” on X appeared to be the first to share the image with the caption: “This is fake”. The person behind the image tells me his name is Blake and that he shared it as part of an experiment. His profile does not have nearly as many followers as the other accounts I have been talking to.

When I ask for evidence that he doctored the image, he told me he has “the original files and creation timestamps”, but he did not share those with me as he says proof does not really matter.

“People share content not because it’s real, but because they want it to be real. Both sides do it equally – they just choose different stories to believe,” he says.

A graphic showing a phone with a screenshot of a post on X by "The Infinite Dude", showing a doctored image of a young Kamala Harris in a McDonald's uniform in front of a large wooden dresser. The image has had a "FALSE" warning pasted over it in red by the BBC.

Blake says he doctored this image to make it appear as though a young Kamala Harris was wearing a McDonald’s uniform

His political allegiance remains unclear and he says this “isn’t about politics”.

Advertisement

X says online that its priority is to protect and defend the user’s voice. The site adds manipulated media labels to some AI-generated and doctored video, audio and images. It also has a feature called Community Notes, which crowdsources fact-checking from users.

During the UK election, X did take action over a network of accounts sharing faked clips that I investigated. In the US election campaign, however, I have received no response to my questions or requests to interview Elon Musk.

That matters – because social media companies like his could affect what unfolds as voters head to the polls.

Marianna Spring investigated this story using her Undercover Voters – five fictional characters based on data from the Pew Research Centre – that allow her to interrogate what some different users are recommended on social media. Their social media accounts are private and do not message real people.

Advertisement

Find out more about them here – and on the BBC Americast podcast on BBC Sounds.

Source link

Continue Reading

Technology

NYT Strands today: hints, spangram and answers for Wednesday, October 30

Published

on

NYT Strands today: hints, spangram and answers for Saturday, September 21

Strands is a brand new daily puzzle from the New York Times. A trickier take on the classic word search, you’ll need a keen eye to solve this puzzle.

Like Wordle, Connections, and the Mini Crossword, Strands can be a bit difficult to solve some days. There’s no shame in needing a little help from time to time. If you’re stuck and need to know the answers to today’s Strands puzzle, check out the solved puzzle below.

How to play Strands

You start every Strands puzzle with the goal of finding the “theme words” hidden in the grid of letters. Manipulate letters by dragging or tapping to craft words; double-tap the final letter to confirm. If you find the correct word, the letters will be highlighted blue and will no longer be selectable.

If you find a word that isn’t a theme word, it still helps! For every three non-theme words you find that are at least four letters long, you’ll get a hint — the letters of one of the theme words will be revealed and you’ll just have to unscramble it.

Advertisement

Every single letter on the grid is used to spell out the theme words and there is no overlap. Every letter will be used once, and only once.

Each puzzle contains one “spangram,” a special theme word (or words) that describe the puzzle’s theme and touches two opposite sides of the board. When you find the spangram, it will be highlighted yellow.

The goal should be to complete the puzzle quickly without using too many hints.

Hint for today’s Strands puzzle

Today’s theme is “How sweet!”

Advertisement

Here’s a hint that might help you: sugary treats.

Today’s Strand answers

NYT Strands logo.
NYT

Today’s spanagram

We’ll start by giving you the spangram, which might help you figure out the theme and solve the rest of the puzzle on your own:

Today’s Strands answers

  • TOFFEE
  • CHOCOLATE
  • GUMDROP
  • LOLLIPOP
  • LICORICE






Source link

Advertisement
Continue Reading

Technology

iPhone 17 early manufacturing reportedly taking place in India

Published

on

iPhone 17 early manufacturing reportedly taking place in India

Apple could take a big step in the process of reducing its dependence on Chinese factories to manufacture iPhones. The company wants to avoid potential supply constraints in the context of the trade war between the United States and the Asian country. A recent report claims that Apple has started the iPhone 17 early manufacturing stage at an Indian factory.

Apple carrying out iPhone 17 early manufacturing stage in Indian factory

Until now, Apple has carried out the entire manufacturing process of iPhone devices in factories in China. However, The Information suggests that this has changed in the early stage of the development of the iPhone 17. This phase is key to determining the viability of mass production of the device.

During the early manufacturing stage, Apple engineers experiment with the potential design, materials, and other key parameters involved in the production of the device. The goal is to ensure that mass manufacturing (millions of devices per day) is possible around the world with as few defective units as possible. Basically, this stage determines the manufacturing bases that will be used for all iPhone 17 units worldwide.

The report specifies that Apple is doing this for the vanilla iPhone 17. This implies that China or other countries will continue to handle the early manufacturing stage for other models in the series. The source claims that it obtained the information from two reliable sources directly related to the situation.

Advertisement

US political context may have driven the decision

The current political context in the United States could have prompted Apple to accelerate the diversification of the production process of its devices. If Trump wins, he aims to apply protectionist economic policies. This includes continuing or escalating the trade war against China with new, severe tariffs. Potential retaliation from Beijing could be a strong blow to companies that depend on factories in the Asian country.

So, Apple seems to want to start the diversification process as soon as possible. The Cupertino giant could be applying early damage control to get ahead of events. If all goes well, Apple could move more stages of the iPhone manufacturing process to Indian factories. In related news, Indonesia banned sales of iPhone 16 devices due to Apple’s failure to comply with local investment.

Source link

Advertisement
Continue Reading

Technology

X is trying to make Community Notes faster with ‘lightning notes’

Published

on

X is trying to make Community Notes faster with ‘lightning notes’

X is trying to speed up its crowdsourced fact-checking system, Community Notes. In an update, the company says it has “re-architected” the scoring system that powers the feature so that the user-generated notes can now appear less than 20 minutes after a post is published on its platform.

Community Notes, introduced in 2022, relies on other X users to fact-check or add missing context to posts on the platform. Contributors are required their sources, and other users then rate the “helpfulness” of the note. Creators are also for posts that get “community noted” in an effort to discourage them from trying to monetize misinformation. Now, that whole process should be able to move a lot quicker.

According to X, these new “lightning notes” can “go live in as little as 14m33s after being written, and 18m20s after the post itself was written.” The change could help address a long running criticism of the crowdsourced fact checking system: that it moves far too slowly compared with the speed of viral misinformation on the platform. For example, last year by Bloomberg found that it could take several hours for a Community Note to appear on a viral tweet and that, often, only a fraction of users see the fact check compared with the original post.

The new speedier system could change that, though it’s unclear how often the faster “lightning” version of the process will actually play out. Not all posts with incorrect information, misstated facts or AI-generated imagery are immediately flagged for review, if they are at all. X says it has more than 800,000 contributors to the program globally, but some posts will likely still take much longer to wind their way through the Community Notes process.

Advertisement

Source link

Continue Reading

Trending

Copyright © 2024 WordupNews.com