FILE PHOTO: Elon Musk, CEO of Tesla, at a convention in Beverly Hills, California, U.S., Could 6, 2024. REUTERS/David Swanson.
Billionaire Elon Musk launched a ballot on social media platform X on Tuesday asking customers whether or not Tesla ought to make investments $5 billion in his artificial intelligence startup xAI.
Board approval and shareholder votes are wanted, Elon Musk stated, including that the ballot is to “check the waters” for a deal.
Tesla didn’t instantly reply to a Reuters request for remark, whereas xAI couldn’t be reached.
As of 0234 GMT, greater than 322,000 customers had participated in the ballot, with 70.2% voting in favour of the funding. The ballot is about to shut in a day.
Musk based xAI final yr as a challenger to Microsoft -backed OpenAI and Alphabet’s Google, which are amongst these main the fierce race for generative AI dominance.
Earlier within the day, Tesla reported its lowest revenue margin in additional than 5 years because the U.S. electrical automobile maker minimize costs to revive demand and elevated spending on AI tasks.
Tesla stated it expects its hardware-related revenue to be accompanied by the acceleration of AI, software program and fleet-based income.
Final month, Musk stated Tesla will probably spend between $3 billion and $4 billion on its purchases of chipmaker Nvidia’s {hardware} this yr.
Of the $10 billion in synthetic intelligence-related capital expenditure this yr, round half can be inside spend, he stated.
xAI, which was valued at $24 billion, had raised in Could $6 billion in sequence B funding, backed by buyers together with Andreessen Horowitz and Sequoia Capital.
xAI has already begun coaching its giant language mannequin (LLM) Grok, on what Musk claims, is the world’s strongest AI coaching cluster positioned in Memphis, Tennessee. It has greater than 1,00,000 Nvidia H100 AI chips.
“It is a vital benefit in coaching the world’s strongest AI by each metric by December this yr,” he stated. Earlier he had stated that “Grok is being skilled in Memphis.”
THIS CONTENT IS BROUGHT TO YOU BY NTNU Norwegian College of Science and Know-how – read more
In an experiment, researchers used synthetic intelligence recreate perfumes. So what would possibly they have the ability to obtain in the future?(Photograph: Colourbox)
Some scents are on the verge of extinction. Nevertheless, with the assist of synthetic intelligence, we would possibly have the ability to deliver again misplaced scents.
It’s a well-known incontrovertible fact that scents can set off our reminiscences. Many of us have smells that remind us of somebody we love, or of Christmas, or of journeys to new or acquainted locations.
However scents may disappear ceaselessly.
For the way did a forest truly smell like earlier than people tampered with it and induced basic adjustments in the composition of its species? What did a new child Tasmanian tiger cub smell like? What about the streets round the Colosseum in ancient Rome or, for that matter, a random metropolis centre on a spring day in 1970?
Recreating fragrances
New expertise will help us protect scents which are disappearing, and synthetic intelligence is a crucial half of the resolution.
“Our analysis makes use of machine studying to recreate olfactory experiences,” says Idelfonso Nogueira.
Idelfonso Nogueira.(Photograph: NTNU)
He’s an affiliate professor at NTNU’s Division of Chemical Engineering. Nogueira is at present engaged on combining new and previous applied sciences to search out options to fashionable issues.
Recreating scents is one of the many areas he’s involved in.
“We had been capable of get synthetic intelligence to recreate fragrances in our experiment,” says Nogueira.
Odours are molecules
However first, what precisely are scents?
Scents are literally totally different combos of chemical compounds in gaseous type. Our olfactory sense, or sense of smell, reacts to the molecules in these chemical compounds. Totally different molecules, and combos of molecules, produce totally different scents.
Scents are influenced not solely by the sorts of molecules concerned but additionally by the ratios between them and the liquids wherein they could be dissolved.
Thus, a selected scent shortly turns into very complicated. It consists of many various components and will be affected by air stress and humidity.
Analysing molecular combos
“We created a mannequin that mixes the molecules in a selected perfume with the human sense of smell,” says Nogueira.
The mannequin used a molecule generator powered by synthetic intelligence. The analysis group used a Gated Graph Neural Community (GGNN), a synthetic neural community that’s skilled to recognise a big quantity of totally different molecules.
This community can interpret the molecular combos of molecules a scent consists of. It will possibly then set up which molecules are concerned, quantify their proportions, and present which molecules are dominant.
This then helps the researchers recreate the scent.
Recreating real-life scents
In a laboratory in Slovenia, a bunch of researchers has recreated scents primarily based on AI recipes. The group is led by Professor Matija Strlič.
“The recreation of scents reveals that this additionally works in apply,” says Nogueira.
Naturally, the expertise could possibly be of nice assist to the fragrance business. Growing perfumes is an advanced course of that entails lots of trial and error. It additionally requires in depth perfume-making expertise.
“Our outcomes are promising, however there’s nonetheless room for enchancment. We have to practice the mannequin on extra molecules and combos with a purpose to recreate scents extra precisely,” he says.
Which means it’s unlikely that synthetic intelligence will take over the jobs of skilled perfumers anytime quickly. As a substitute, it could actually develop into a useful instrument.
Perfumers aren’t the solely ones who can profit from this expertise. In the future, it’d assist us to protect different scents as nicely.
Maybe in the future it will likely be potential to recreate the scent of a beloved member of the family, a childhood residence, or a primeval forest?
For now, a lot of the work is carried out in laboratories at NTNU. Nevertheless, researchers have now obtained help to collaborate with establishments in Italy, Latvia, Slovenia, the Czech Republic, and the UK.
Search. Cloud. AI. It was a profitable mixture for Google’s dad or mum Alphabet in Q1 and it labored once more in Q2, as reported in the corporate’s earnings name after the bell on Tuesday (July 23).
By the numbers, year-over-year web earnings in the quarter ending June 30 rose 28.6% to $23.6 billion, beating the typical Wall Road estimate of $22.9 billion. Total income from all product strains, together with Android and YouTube, grew 14% to $84.74 billion. Promoting gross sales, Alphabet’s chief income supply, rose 11% to $64.6 billion. Income from cloud computing providers, a sign of enterprise expertise spending, rose 28.8% to $10.35 billion.
However the actual star of the present was synthetic intelligence (AI). In line with Alphabet executives on the decision, it should almost definitely take the lion’s share of the $12 billion per quarter anticipated from the corporate’s capital expenditure funds. It has already generated “billions” from its prime 100 enterprise clients, though precise numbers weren’t forthcoming. And by combining AI with conventional search it’s attracting the crucial 18-24-year-old buyer phase.
Whereas different corporations like Apple and Meta are partnering for AI mannequin growth, Alphabet CEO Sundar Pichai put his technique on blast, saying it should enable it to manage its personal future for a expertise he expects to affect each product and line of enterprise the corporate has.
“Our analysis and infrastructure management means we will pursue an in-house technique that permits our product groups to maneuver rapidly,” Pichai informed the decision. “Mixed with our model-building experience, we’re in a powerful place to manage our future because the expertise continues to evolve. Importantly, we are innovating at each layer of AI stack, from chips to brokers and past. An enormous energy. We’re dedicated to this management lengthy time period.”
Pichai detailed a number of AI-driven improvements, together with visible search through Google Lens and the upcoming characteristic that enables customers to ask questions by video. The “circle to go looking” characteristic, now out there on over 100 million Android gadgets, exemplifies the momentum gained from Google’s AI investments. Moreover, the Gemini AI mannequin household is powering enhancements throughout Google’s suite of merchandise, from search and workspace to Google Images and Google Messages.
“Our AI product advances come from our long-standing basis of analysis management and our international community of infrastructure,” Pichai emphasised, highlighting the institution of latest information facilities and cloud areas, comparable to the primary information middle in Malaysia and expansions in Iowa, Virginia and Ohio. The introduction of the sixth era Trillium AI accelerator, which provides vital enhancements in efficiency and power effectivity, additional solidifies Google’s infrastructure capabilities.
YouTube’s income and engagement metrics have been a focus of the earnings name, with sturdy development reported in each adverts and subscriptions. Philipp Schindler, chief enterprise officer, shared that YouTube’s advert income elevated by 13% yr over yr, pushed by development in model and direct response promoting. “YouTube has remained primary in U.S. streaming watch time,” Schindler famous, with views of YouTube Shorts on related TVs greater than doubling over the previous yr.
Subscriptions have turn out to be a crucial income stream for YouTube, contributing to the corporate’s monetary efficiency. “12 months-on-year revenues elevated 14%, pushed once more by sturdy development in YouTube subscriptions,” Schindler stated. The introduction of new options, comparable to simpler methods for creators so as to add captions and convert common movies into Shorts, has bolstered consumer engagement and content material creation.
Schindler additionally highlighted the strategic initiatives aimed toward enhancing the promoting expertise by AI. Google’s efforts to include AI into the advertising course of have led to the event of over 30 new advert options and merchandise. These improvements are designed to streamline workflows, improve artistic asset manufacturing, and supply extra participating shopper experiences. Retailers, in explicit, he stated, have benefited from AI-powered instruments that scale the depth and high quality of their promoting property.
For all PYMNTS AI protection, subscribe to the day by day AI Newsletter.
