28.7 C
Sunday, July 21, 2024

Who is Sam Altman? The former OpenAI boss and ChatGPT guru who became one of AI’s biggest players | Science & Tech News

Must read

He was a tech whizz before he left primary school, dropped out of one of America’s top universities, and appeared to be spearheading a revolution that could change our lives forever.

Sam Altman would have been unknown to most outside tech circles before the launch of his firm’s breakthrough chatbot ChatGPT, but he has recently much of his time rubbing shoulders with world leaders and some of America’s most recognisable executives.

But in a surprise announcement on Friday, OpenAI – the firm behind ChatGPT – revealed Altman had been ousted as its chief executive after the board said it no longer had confidence in him.

Here, Sky News looks at the 38-year-old’s rise to fame – before his sudden axing.

Earlier life

Altman grew up in the US state of Missouri where, as an eight-year-old, he was gifted his first computer and quickly learnt not just how to use it, but to program for it.

Altman attended John Burroughs School in St Louis, and told The New Yorker in a 2016 interview that having his computer helped him come to terms with his sexuality and come out to his parents when he was a teenager.

More on Artificial Intelligence

“Growing up gay in the Midwest in the 2000s was not the most awesome thing,” he recalled. “And finding AOL chatrooms was transformative. Secrets are bad when you are 11 or 12.”

With school in the rear-view mirror, it was time for university – Stanford, no less. Altman made his way to that famous California institution to study computer science, but dropped out after just two years, following in the footsteps of previous dropouts-turned-tech superstars Bill Gates and Mark Zuckerberg, who both abandoned their Harvard degrees before becoming two of history’s most influential CEOs.

Abandoning a precious spot at one of America’s top universities seemed such a rite of passage for the country’s leading tech entrepreneurs that it played right into the success story of the now disgraced Elizabeth Holmes, whose departure from Stanford to gatecrash Silicon Valley led to a wave of media attention not dissimilar to that currently given to Altman.

His first post-university venture was a smartphone app called Loopt, which let users selectively share their real-time location with other people. Some $30m (£24m) was raised to launch the company, aided by funding from a start-up accelerator firm called Y Combinator, which lists the likes of Airbnb and Twitch among the internet companies it has helped establish.

Altman became president of Y Combinator itself in 2014, after the sale of Loopt for $44m (£35m) in 2012. He also founded his own venture capital fund called Hydrazine Capital, attracting enough investment to be named on the Forbes 30 Under 30 list for venture capital. As if he wasn’t busy enough, Altman also ran Reddit for a grand total of eight days amid a leadership shake-up in 2014, describing his tenure as “sort of fun”.

The rise of OpenAI

While his time at the top of Reddit only lasted eight days, his oversight of OpenAI has lasted eight years. He was “doing pretty well” with it, he said in a February tweet (certainly compared to Loopt, which, he now says, “sucked”).

He launched the company with a certain Elon Musk (who only ran SpaceX and Tesla at the time) in 2015, the two men providing funding alongside the likes of Amazon and Microsoft, totalling $1bn (£800m).

It was run as a non-profit with the noble goal of developing AI while making sure it doesn’t wipe out humanity.

So far, mission accomplished – but if Altman’s to be believed, the risk since has become very real indeed.

Under his tenure, OpenAI ceased to be a non-profit and grew to an estimated value of up to $29bn (£23bn), all thanks to the remarkable success of its generative AI tools – ChatGPT for text and DALL-E for images.

Microsoft boss Satya Nadella described Altman as an “unbelievable entrepreneur” who bets big and bets right, which OpenAI’s success makes hard to argue with.

ChatGPT amassed tens of millions of users within weeks of launching in late 2022, wowing experts and casual observers alike with its ability to pass the world’s toughest exams, get through job applications, compose anything from political speeches to children’s homework, and write its own computer code.

Suddenly the concept of a large language model (meaning it is trained on huge amounts of text data so that it can understand our requests and respond accordingly) became something of a mainstream buzz term, its popularity seeing Microsoft invest extra cash into OpenAI and bring the tech to its Bing search engine and Office apps.

Google also got in on the act with its Bard chatbot, some of China’s biggest tech companies entered the race, while Musk – who left OpenAI in 2018 due to a conflict of interest with Tesla’s work on self-driving AI – has said he wants to launch his own one too.

All the while, OpenAI’s technology is also improving – an upgrade dubbed GPT-4 within months of ChatGPT’s release showing just how quickly these models can develop.

Read more:
We asked a chatbot to help write an article

Please use Chrome browser for a more accessible video player

Will this chatbot replace humans?

‘My worst fears’

For all the wonder such systems have provided, it’s matched – if not surpassed – by the concerns. Whether it be spreading disinformation or making jobs redundant, governments are scrambling to formulate an effective way of regulating a technology that seems destined to change the world forever.

Perhaps with an eye on how some of his Silicon Valley contemporaries have failed to act on the dangers of their creations before it’s too late, Altman appears keen to be a willing participant in just how it should be done.

“My worst fears are that we, the industry, cause significant harm to the world,” Altman told the US Senate, his assessment that government regulation would be “critical to mitigate the risks” undoubtedly music to the ears of politicians who never seem overly impressed by figures from the tech world.

Read more:
Who is the ‘godfather of AI’?

Please use Chrome browser for a more accessible video player

AI speech used to open Congress hearing

In the space of a few short weeks, Altman met the US vice president, Kamala Harris, France’s Emmanuel Macron, European Commission president Ursula von der Leyen and the British prime minister, Rishi Sunak – all politicians who share the same hopes and fears about the potential benefits and dangers of AI.

With the EU seemingly none too impressed by Elon Musk’s running of Twitter, TikTok managing to achieve the mostly impossible task of uniting Democrats and Republicans against a common enemy, and Mark Zuckerberg having struggled to repair his reputation after the Cambridge Analytica scandal, the upstart Altman could be positioning himself to become a more durable tech star than some of his forebears.

Prime Minister Rishi Sunak met the CEOs of OpenAI, Google DeepMind, and Anthropic
Sam Altman meets with PM Rishi Sunak in May 2023

But just in case it does all go wrong, he’s previously admitted to being a prepper – someone who stockpiles everything from guns to medicine should the worst should befall us.

Announcing Altman’s departure as OpenAI chief executive on, the company said a review found he had not been “consistently candid in his communications with the board”.

He posted on social media following the announcement, writing: “I loved my time at OpenAI.

“It was transformative for me personally, and hopefully the world a little bit.

“Most of all, I loved working with such talented people. (I) will have more to say about what’s next later.”

Source link

- Advertisement -spot_img

More articles

- Advertisement -

Latest articles