OpenAI's long-awaited GPT-5 model nears release
Two early testers of the new model told Reuters they have been impressed with its ability to code and solve science and math problems, but they believe the leap from GPT-4 to GPT-5 is not as large as the one from GPT-3 to GPT-4. The testers, who have signed non-disclosure agreements, declined to be named for this story.
OpenAI declined to comment for this story.
GPT-4's leap was based on more compute power and data, and the company was hoping that 'scaling up' in a similar way would consistently lead to improved AI models.
But OpenAI, which is backed by Microsoft (MSFT.O), opens new tab and is currently valued at $300 billion, ran into issues scaling up. One problem was the data wall the company ran into, and OpenAI's former chief scientist Ilya Sutskever said last year that while processing power was growing, the amount of data was not.
He was referring to the fact that large language models are trained on massive datasets that scrape the entire internet, and AI labs have no other options for large troves of human-generated textual data.
Apart from the lack of data, another problem was that 'training runs' for large models are more likely to have hardware-induced failures given how complicated the system is, and researchers may not know the eventual performance of the models until the end of the run, which can take months.
OpenAI has not said when GPT-5 will be released, but the industry expects it to be any day now, according to media reports. Boris Power, head of Applied Research at OpenAI, said in an X post on Monday: "Excited to see how the public receives GPT-5."
'OpenAI made such a great leap from GPT-3 to GPT-4, that ever since then, there has been an enormous amount of anticipation over GPT-5,' said Navin Chaddha, managing partner at venture capital fund Mayfield, who invests in AI companies but is not an OpenAI investor. 'The hope is that GPT-5 will unlock AI applications that move beyond chat into fully autonomous task execution."
Nearly three years ago, ChatGPT introduced the world to generative AI, dazzling users with its ability to write humanlike prose and poetry, quickly becoming one of the fastest growing apps ever.
In March 2023, OpenAI followed up ChatGPT with the release of GPT-4, a large language model that made huge leaps forward in intelligence. While GPT-3.5, an earlier version of the model, received a bar exam score in the bottom 10%, GPT-4 passed, opens new tab the simulated bar exam in the top 10%.
GPT-4 then became the model to beat and the world came to terms with the fact that AI models could outperform humans in many tasks.
Soon, other companies were catching on. The same year, Alphabet's Google (GOOGL.O), opens new tab and Anthropic - which is backed by Amazon (AMZN.O), opens new tab and Google - released competitive models to GPT-4. Within a year, open-source models on par with GPT-4 such as Meta Platforms' (META.O), opens new tab Llama 3 models were released.
Along with training large models, OpenAI has now invested in another route, called 'test-time compute,' which channels more processing power to solve challenging tasks such as math or complex operations that demand human-like reasoning and decision-making.
The company's CEO Sam Altman said earlier this year that GPT-5 would combine both test-time compute and its large models. He also said that OpenAI's model and product offerings had become "complicated."
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Independent
18 minutes ago
- The Independent
Trump's planned 100% computer chip tariff sparks confusion among businesses and trading partners
President Donald Trump 's plans for 100% tariffs on computer chips that aren't made in the U.S. are stoking confusion among businesses and trading partners — boosting stocks for leading semiconductor companies while leaving smaller producers scrambling to understand the implications. The U.S. imports a relatively small number of chips because most of the foreign-made chips in a device — from an iPhone to a car — were already assembled into a product, or part of a product, before it landed in the country. "The real question everybody in the industry is asking is whether there will be a component tariff, where the chips in a device would require some sort of separate tariff calculation,' said Martin Chorzempa, a senior fellow at the Peterson Institute for International Economics. Trump said Wednesday that companies that "made a commitment to build" in the U.S. would be spared the import tax, even if they are not yet producing those chips in American factories. 'We'll be putting a tariff of approximately 100% on chips and semiconductors,' Trump said in the Oval Office while meeting with Apple CEO Tim Cook. 'But if you're building in the United States of America, there's no charge.' Wall Street investors interpreted that as good news not just for U.S. companies like AMD, Intel and Nvidia, but also for the biggest Asian chipmakers like Samsung and Taiwan Semiconductor Manufacturing Company that have been working to build U.S. factories. But it left greater uncertainty for smaller chipmakers in Europe and Asia that have little exposure to the AI boom but still make semiconductors inserted into essential products like cars or washing machines. These producers "probably aren't large enough to get on the map for an exemption and quite probably wouldn't have the kind of excess capital and margins to be able to add investment at a large scale into the United States,' Chorzempa said. The announcement came more than three months after Trump temporarily exempted most electronics from his administration's most onerous tariffs. During the COVID-19 pandemic, a shortage of computer chips increased the price of autos and contributed to higher inflation. Chorzempa said chip tariffs could again raise prices by hundreds of dollars per vehicle if the semiconductors inside a car are not exempt. 'There's a chip that allows you to open and close the window," Chorzempa said. "There's a chip that is running the entertainment system. There is a chip that's kind of running all the electronics. There are chips, especially in EVs, that are doing power management, all that kind of stuff.' Much of the investment into building U.S. chip factories began with the bipartisan CHIPS and Science Act that President Joe Biden signed into law in 2022, providing more than $50 billion to support new computer chip plants, fund research and train workers for the industry. Trump has vocally opposed those financial incentives and taken a different approach, betting that the threat of dramatically higher chip costs would force most companies to open factories domestically, despite the risk that tariffs could squeeze corporate profits and push up prices for electronics.
The Independent
18 minutes ago
- The Independent
The species that can regrow their eyes after amputation
Scientists are studying the freshwater apple snail's ability to fully regenerate its eyes, hoping to apply this knowledge to help humans with eye injuries. Research has revealed that the complex structure of human eyes shares significant anatomical and genetic similarities with those of the apple snail. Led by Alice Accorsi at the University of California, the team is using genome editing techniques like CRISPR-Cas9 to understand the genetic and molecular mechanisms behind the snail's eye regeneration. Apple snails are ideal for this research due to their resilience, rapid breeding, and 'camera-type' eyes, which are similar to human eyes and produce high-resolution images. The study details the snail's regeneration process, showing how new eye structures, including the lens and retina, form within weeks after amputation, with the ultimate aim of identifying genes that could enable human eye regeneration.
Reuters
19 minutes ago
- Reuters
US jobless claims edge up, but 'no-hire, no-fire' trend remains intact
Aug 7 (Reuters) - The number of Americans filing new applications for unemployment benefits ticked up to the highest level in a month last week, suggesting the labor market was largely stable even though job creation is weakening and it is taking laid-off workers longer to find new employment. Initial claims for state unemployment benefits for the week ended August 2 rose 7,000 to a seasonally adjusted 226,000, the highest level since the week ending July 5, the Labor Department said on Thursday. Economists polled by Reuters had forecast 221,000 claims for the latest week. The labor market has slowed, with government data last week showing far fewer jobs were created in July than economists had expected as uncertainty over President Donald Trump's tariffs left businesses wary of adding workers. Moreover, employment gains in the previous two months were revised lower by nearly 260,000, a stunning reversal that prompted Trump to fire the head of the Bureau of Labor Statistics - a move that rattled investors and economists already anxious about the eroding quality of official U.S. economic data. The latest data on new claims indicates employers are not yet turning to large-scale layoffs as the economy loses steam but are managing through attrition. That has helped keep the unemployment rate, at 4.2% in July, relatively low even while job growth has slowed. Declining labor supply amid the White House's immigration crackdown is also helping to stave off a jump in the jobless rate. Employers' hesitancy to increase hiring means there are fewer jobs for those being laid off. The number of people receiving benefits after an initial week of aid, a proxy for hiring, rose to a seasonally adjusted 1.974 million during the week ending July 26, the claims report showed. That was the highest level of so-called continued claims since November 2021, and the increase of 38,000 from the previous week was the largest since late May. Despite the rise in both new and ongoing claims last week, economists note the two continue to remain largely range-bound and have not exhibited the kind of breakout upswing that in the past has signaled a deteriorating job market. Indeed, the weekly net change in new claims has remained below 10,000 - in either direction - for 10 straight weeks, the longest such streak in about three years. "The sideways drift in initial and continuing claims in recent months suggests that layoff activity is muted," Thomas Simons, chief U.S. economist at Jefferies, wrote in a note. "The 'no hire/no fire' theme in the labor market remains firmly intact." That said, continued claims had been averaging less than 1.9 million until early May and in the three months since have averaged about 40,000 higher at about 1.94 million. That reset now makes more sense to some economists in light of the historic downward revision to the job creation figures for May and June that were revealed by BLS last week. "The level of continued claims is still signaling that unemployed workers are finding it tough to find a job in a labor market where hiring is slow," Nancy Vanden Houten, lead U.S. economist at Oxford Economics, wrote. "Moreover, the steep rise in claims since April makes more sense following the sharp downward revisions to job growth in May and June and the sluggish pace of hiring in July." Meanwhile, the Labor Department also said worker productivity rebounded more than expected in the second quarter, easing a surge in labor costs at the start of the year. Nonfarm business sector productivity increased 2.4% in the April-June period after declining by 1.8% in the first three months of the year, BLS reported. Economists polled by Reuters had expected productivity to rise by 2.0%. Worker output increased by 3.7%, the largest improvement since the third quarter of 2023, while unit labor cost growth moderated to 1.6% from an upwardly revised 6.9% in the first quarter. The data may be indicative of ongoing investments by businesses in labor-saving technologies like artificial intelligence, economists said. "Productivity growth is settling back into its historical trend after bouncing around because of the pandemic," said Oren Klachkin, financial market economist at Nationwide. "Looking ahead, businesses will likely invest in labor-saving technologies to cap their wage bills, which should exert downward pressure on inflation. Companies facing the greatest labor constraints will likely be the most incentivized to invest."
