The Atlanta Journal-Constitution
AI company CEO is tired of doomsday allusions
Aidan Gomez can take some credit for ‘T’ at the end of ChatGPT.
Aidan Gomez can take some credit for the “T” at the end of ChatGPT. He was part of a group of Google engineers who introduced an artificial intelligence model called a transformer. That helped set a foundation for today’s generative AI boom that ChatGPT maker OpenAI and others built upon.
Gomez, one of eight co-authors of Google’s 2017 paper, was a 20-year-old intern at the time. He’s now the CEO and co-founder of Cohere, a Toronto-based startup competing with other AI companies in supplying to big businesses and organizations large language models and the chatbots they power.
Gomez spoke with The Associated Press about the future of generative AI. The interview has been edited for length and clarity.
Q: What’s a transformer?
A: A transformer is an architecture of a neural network — the structure to the computation that happens inside of the model. The reason that transformers are special relative to their peers — other competing architectures, other ways of structuring neural networks — is essentially that they scale very well. They can be trained across not just thousands but tens of thousands of chips. They can be trained extremely quickly. They use many different operations that these GPUs (graphics chips) are tailored for. Compared to what existed before the transformer, they do that processing faster and more efficiently. conversation and pre-populating those notes so that instead of having to write it from scratch, there’s a first draft in there. They can read through it and just make edits. Suddenly, the capacity of doctors boosts by a massive proportion.
Q: How do you address customer concerns about AI language models being prone to errors and bias?
A: Customers are always concerned about “hallucinations” (errors) and bias. It leads to a bad product experience. So it’s something we focus on heavily.
For hallucinations, we have a core focus on RAG, which is retrieval-augmented generation. We just released a new model called Command R, which is targeted explicitly at RAG. It lets you connect the model to private sources of trusted knowledge. That might be your organization’s internal documents or a specific employee’s emails. You’re giving the model access to information that it just otherwise hasn’t seen out in the web when it was learning.
What’s important is that it also allows you to factcheck the model because now instead of just text in, text out, the model is actually making reference to documents. It can cite back to where it got that information. You can check its work and gain a lot more confidence working with the tool. good.
A lot of the arguments for how it might go bad — those of us developing the technology are aware of and working to mitigate those risks. We all want this to go well. We all want the technology to be additive to humanity, not a threat to it.
Q: Not only OpenAI but a number of major technology companies are now explicitly saying they’re trying to build “artificial general intelligence” (a term for broadly better-than-human AI). Is AGI part of your mission?
A: No, I don’t see it as part of my mission. For me, AGI isn’t the end goal. The end goal is profound positive impact for the world with this technology. It’s a very general technology. It’s reasoning, it’s intelligence, so it applies all over the place.
We want to make sure it’s the most effective form of the technology it possibly can be, as early as it possibly can be. It’s not some pseudo-religious pursuit of AGI, which we don’t even really know the definition of.
Sunday, March 31, 2024 |
■ “Take What You Need,” by Idra Novey. (Penguin, 256 pages, $18.) When her former stepmother Jean dies and leaves Leah the sculptures she spent her final years creating, Leah returns to Appalachia, where the two once lived together. Contending with “ambivalence and complexity and discomfort,” a New York Times reviewer wrote, this novel traces their relationship over the course of Leah’s life and across divides of generation and class.
Beauty trends come and go, but one thing has been a makeup mainstay for millennia: deep red lipstick. Women – and, yes, men too – have painted their faces with this hue since the start of civilization.
Now, archaeologists have identified the oldest known physical example of lipstick. Carbon dated to between 1936 and 1687 B.C., the analyzed contents of a small stone vial from Iran are a Bronze Age form of dark red lipstick that was probably applied with a brush.
“It is important to note that in the present state of knowledge, we hypothesize a lippaint, rather than a solid lipstick,” Massimo Vidale, one of seven scientists who studied the substance, said in an email.
Looking very much like a modern tube of lipstick, the cylindrical container carrying the suspected lip paint was unearthed in 2001 from a 3rd millennium B.C. graveyard near the Halil River in Kerman Province, in southeastern Iran. Vidale and his team of researchers recently examined the container and published their findings last month in Scientific Reports.
“It is always a sudden revelation,” said Vidale, a professor of archaeology at the University of Padua in Italy. “Even more, because the substance under our eyes matched perfectly with the very unusual form of the fine stone container, so close to the lipstick cases of our days.”
It’s ironic that the first known lipstick specimen was found in Iran, a country that banned cosmetics like lipstick and nail polish following the 1979 Islamic revolution. Though restrictions have been relaxed, many women – including government workers and medical students – are still prohibited from wearing makeup.
Vidale said the vial’s contents were in powder form due to dehydration after more than 3,700 years. Under the microscope, dark pigments were clearly visible, indicating the original product was an “intense red color,” the team reported in the study.
Using an environmental scanning electron microscope, the scientists were able to identify the makeup’s makeup. It features a large amount of hematite, a ferric oxide mineral whose name is derived from the Greek word for “blood” because of its red color. The contents also include vegetable oils and waxes, which are found in today’s lipstick as well.
“Combined together, this is exactly what one would expect in a modern lipstick,” Vidale wrote.
The use of lipstick dates back more than 5,500 years. According to ancient texts, Queen Puabi (also known as Shub-Ad) of Ur in Mesopotamia was the first to apply coloring to her mouth.
The Sumerian sovereign apparently used a mixture of white lead and crushed red rocks around 3500 B.C. to rouge her lips, according to a 2006 paper by Sarah Schaffer, published by Harvard University.
How that toxic cosmetic affected the queen’s health is not known since much of her life remains a mystery.
Ancient Egyptians were also known to use lipstick for ceremonial, funerary and everyday purposes. A famous illustration from around 1150 B.C. shows a woman anointing her lips with what could be makeup. Known as Turin Papyrus 55001, the drawing depicts a young Egyptian with a brush in one hand while the other holds a mirror and what appears to be “a thin, round-bottomed cylindrical cosmetic vial,” similar in size and shape to the Iranian container, according to the Scientific Reports article.
Vidale and his fellow researchers, whose study of the lipstick was supported by the universities of Tehran and Padua and the International Association for Mediterranean and Oriental Studies in Rome, became interested in the vial when team member Nasir Eskandari spotted it in the collection of the Jiroft Archaeological Museum in Iran.
At the time, its purpose was unknown. The container was originally unearthed when the Halil River flooded ancient graveyards in 2001 and exposed numerous artifacts. Many items were looted but then recovered by Iranian security forces.
Measuring approximately 2 inches tall by three-quarters of an inch wide, the stone vial is slightly smaller than a modern tube of lipstick.
The hand-carved container is made of chlorite and includes a hole on the top for extracting the contents. Its chiseled features give it the appearance of a marsh reed – offering a clue to its purpose.
“Because of its shape, mimicking a segment of a marsh cane, it is very thin and easy to handle,” Vidale told The Washington Post. “At the time, people used common cane segments as cases or receptacles, but this one is made of a prestigious stone.”
The vial is believed to have been used by the Jiroft culture. This Bronze Age civilization was unknown until the 2001 graveyard flooding. Its capital is thought to have been the city-state of Marhaši, which is mentioned in ancient texts but has yet to be located.
“What we know today is that this was an advanced Mesopotamian-like civilization, a major player in long-distance trade and military ventures, which used its own writing system and was ruled by large cities and powerful, authoritative rulers,” Vidale said. “The rest is slowly emerging from new excavations.”
Exactly how this lipstick was used and by whom is not known. Since the vial was buried in a grave, Vidale speculated that it may have been part of funeral rituals.
“We are groping in the dark here,” he wrote. “We can assume that the dead were adorned for their final journey, but also that cosmetics were important personal possessions of daily use, and that, upon an individual’s death, they could not be used by others.”
The discovery demonstrates to Vidale and his team how closely modern society resembles so-called primitive civilizations of long ago.
“The evidence supports a simple point: the world of the early Bronze age, when the first cities became the political hubs of powerful and hierarchic states, was already very, very similar to our one,” he wrote.