 |
|
|
|||||
Anthropic, Menlo Ventures, and other AI industry players are betting $50 million on a company called Goodfire, which aims to understand how AI models think and steer them toward better, safer answers. Even as AI becomes more embedded in business systems and personal lives, researchers still lack a clear understanding of how AI models generate their output. So far, the go-to method for improving AI behavior has focused on shaping training data and refining prompting methods, rather than addressing the models internal thought processes. Goodfire is tackling the latterand showing real promise. The company boasts a kind of dream team of mechanistic interpretability pioneers. Cofounder Tom McGrath helped create the interpretability team at DeepMind. Cofounder Lee Sharkey pioneered the use of sparse autoencoders in language models. Nick Cammarata started the interpretability team at OpenAI alongside Chris Olah, who later cofounded Anthropic. Collectively, these researchers have delivered some of the fields biggest breakthroughs. Goodfire founder and CEO Eric Ho, who left a successful AI app company in 2022 to focus on interpretability, tells Fast Company that the new funding will be used to expand the research team and enhance its Ember interpretability platform. In addition to its core research efforts, Goodfire also generates revenue by deploying field teams to help client organizations understand and control the outputs of their AI models. Goodfire is developing the knowledge and tools needed to perform brain surgery on AI models. Its researchers have found ways to isolate modules within neural networks to reveal the AIs thoughts. Using a technique they call neural programming, they can intervene and redirect a models cognition toward higher-quality, more aligned outputs. We envision a future where you can bring a little bit of the engineering back to neural networks, Ho says. The company has also been collaborating with other AI labs to solve interpretability challenges. For example, Goodfire has helped the Arc Institute interpret the inner workings of its Evo 2 DNA foundation model, which analyzes nucleotide sequences and predicts what comes next. By understanding how the model makes its predictions, researchers have uncovered unique biological conceptspotentially valuable for new scientific discoveries. Anthropic, too, may benefit from Goodfires insights. “Our investment in Goodfire reflects our belief that mechanistic interpretability is among the best bets to help us transform black-box neural networks into understandable, steerable systemsa critical foundation for the responsible development of powerful AI,” Anthropic CEO Dario Amodei said in a statement. According to Ho, Goodfire has also been fielding requests from Fortune 500 companies that want to better understand how the large language models they use for business are thinkingand how to change faulty reasoning into sound decision-making. He notes that many within businesses still see AI models as another kind of software, something that can be reprogrammed when it produces incorrect outputs. But AI works differently: It generates responses based on probabilities and a degree of randomness. Improving those outputs requires intervention within the models cognitive processes, steering them in more productive directions. This kind of intervention is still a new and imprecise science. It remains crude and at a high level and not precise, Ho says. Still, Goodfire offers an initial tool kit that gives enterprises a level of control more familiar from traditional deterministic software. As companies increasingly rely on AI for decisions that affect real lives, Ho believes the ability to understand and redirect AI models will become essential. For instance, if a developer equips a model with ethical or safety guardrails, an organization should be able to locate the layer or parameter in the neural network where the model chose to bypass the rulesor tried to appear compliant while it wasnt. This would mean turning the AI black box into a glass box, with tools to reach inside and make necessary adjustments. Ho is optimistic that interpretability research can rise to the challenge. This is a solvable, tractable, technical problem, but it’s going to take our smartest researchers and engineers to solve the really hard problem of understanding and aligning models to human goals and morals. As AI systems begin to surpass human intelligence, concerns are growing about their alignment with human values and interests. A major part of the challenge lies in simply understanding whats happening inside AI models, which often think in alien, opaque ways. Whether the big AI labs are investing enough in interpretability remains an open questionone with serious implications for our readiness for an AI-driven future. Thats why its encouraging to see major industry players putting real funding behind an interpretability research lab. Lightspeed Venture Partners, B Capital, Work-Bench, Wing, and South Park Commons also participated in the funding round. Menlo Ventures partner Deedy Das will join Goodfires board of directors. While most of the tech world now rushes ahead with the development and application of generative AI models, concerns about the inscrutable nature of the models often get brushed aside as afterthoughts. But that wasnt always the case. Google hesitated to put generative models into production because it feared being sued over unexpected and unexplainable model outputs. In some industries, however, such concerns remain very relevant, Das points out. There are extremely sensitive use cases in law, finance, and so on, where trying to deploy AI models as we know them today is just not feasible because you’re relying on a black box to make decisions that you don’t understand why it’s making those decisions, Das says. A good part of [Goodfires] mission is just to be able to do that.
Category:
E-Commerce
These days, when you head to a shop to buy clothes, most brands package your purchases in a recyclable paper bag, which looks more eco-friendly than plastic. But behind the scenesin back rooms that most customers never seeevery single clothing retailer has enormous piles of flimsy plastic bags (sometimes called poly bags). These bags keep clothes clean as they travel across the complex global supply chain before arriving at the store. We need to keep clothes in good condition as they move from factories to shipping containers to trucks, says Candan Erenguc, chief operations officer at Anthropologie. [Photo: WM/Anthropologie] Most local recycling facilities don’t have the equipment to recycle poly bags, which are more complicated to break down than more solid plastics like water bottles. So most retailers simply send them in the regular waste stream where they will end up in a landfill. Since plastic does not biodegrade, these bags will break down into tiny fragments of microplastic that will end up in our waterways and food. Anthropologie has been on a mission to find a way to recycle the poly bags it collects across its 215 retail stores. Over the past 18 months, it has partnered with Waste Management (WM), the largest recycling company in the United States, to develop a solution. Now, store associates collect these bags and send them to special facilities that are equipped to recycle them into other plastic products, extending their life. Anthropologie has already recycled more than 60,000 tons of poly bags, which have been transformed into pellets that will be used to create other plastic items, including trash bags. It has been a very seamless process, and we want to make sure other retailers know they can do it as well, says Erenguc. That said, things like trash bags cannot be further recycled, so they will eventually end up in a landfill. So it is still incumbent on brands to find ways to reduce the amount of plastic they consume and discard. For decades, flimsy plastic bags have been a challenge for municipal recycling facilities that collect household waste. If you accidentally put them in your curbside recycling bin, they can clog up the recycling equipment, shutting the system down. As a result, people have been encouraged to simply dispose of these bags in the regular waste stream, where they will be landfilled or incinerated. However, recycling technology is quickly improving, according to Tara Hemmer, chief sustainability officer at WM. For one thing, WM is now investing in robotics and computer vision technology that can better catch plastic bags that end up in the waste stream and separate them from the rest of the trash, so they don’t cause a major disturbance. And perhaps more impressively, there are now several industrial recycling facilities across the U.S. that are specifically designed to recycle poly bags. Some of these plants are owned by WM. But there are also independent recyclers that partner with WM. We work with our customers to make sure they can direct their waste to the right facility in our third-party network, says Hemmer. [Photo: WM] Erenguc wanted to find a way to collect poly bags and ship them to these locations. However, as a major retailer, this presented a logistical challenge. It was also important for the process to be easy for employees to understand and follow. Each of Anthropologie’s 215 stores is staffed with dozens of employees who must be trained on best practices when it comes to waste disposal. Moreover, it was unclear where the nearest recycling facility would be for each store. We didn’t want to be transporting poly bags back and forth across the country, because that isn’t good for the environment either, Erenguc says. But this is where WM could help. Anthropologie brought in members of the WM team to study the situation and come up with a solution that would be easy for retail employees to adopt. WM identified the address of the closest recycling facility for each store. Retail associates now collect plastic bags and when they have achieved a certain volume, they ship them out to a designated facility. The recycling plants turn poly bags into pellets that can than be used to create other products. It’s such a streamlined solution, Erenguc says. It was so easy to execute, but we’ve already managed to divert 60,000 pounds of plastic from landfills. [Photo: WM] Hemmer says that many retailers are eager to divert waste from landfill. While there’s been a narrative that companies have abandoned their sustainability goals, that hasn’t been her experience. We’ve found that companies still have goals and are marching towards them, she says. And consumerproduct companies are trying to increase the amount of recycled content that goes into their products. Hemmer says that recycling technology is improving every year. WM is currently working to make it possible to recycle plastic bags in residential areas, beginning with a plant in Chicago that will reach about 3,500 households. But often the obstacle to bringing about change at scale isn’t technologicalit’s logistical. People, as well as companies, are more likely to adopt new processes if they’re simple. Part of our job is to help troubleshoot, says Hemmer. But diverting waste from landfill is actually a lot easier than you’d imagine.
Category:
E-Commerce
Elon Musk’s foray into government has proven disastrous for his business life. Since taking up work for President Donald Trumps’ so-called Department of Government Efficiency (DOGE), Musk’s electric car company Tesla has seen sales slide and has become a target for protests. Now some believe that damage could be terminal and that Musk poses a risk to companies outside of his own. The Reputation Risk Index looks at reputational threats facing companies and organizations. It recently found that being associated with Musk posed the second biggest threat to companies, between the harmful or deceptive use of artificial intelligence and backtracking on DEI. The index, which is based on a survey with 117 public affairs leaders and former heads of state, found it’s not just being associated with Musk that’s risky, but being singled out and publicly criticized by him. In an aerial view, brand-new Tesla cars and Cybertrucks sit parked in a lot at a Tesla dealership on April 02, 2025, in Corte Madera, California. [Photo: Justin Sullivan/Getty Images] With his controversial omnipresence in the media landscape, 28% of the council identified this association as a top reputational risk, highlighting Musks impact on businesses that extend well past his own, Global Risk Advisory Council chair Isabel Casillas Guzman said in the report. Wedbush Securities analyst Dan Ives predicted in a note Sunday that even if Musk were to quit DOGE and get back to his car company there will be permanent brand damage.” And if Musk stays in government, brand damage could grow for Tesla, calling it a code red situation for the company. Musk “needs to leave the government, take a major step back on DOGE, and get back to being CEO of Tesla full-time,” Ives wrote. Musk’s hard turn to DOGE has shown that mixing business with politics can backfire, especially for a public CEO of a company that relies on customers who in large part don’t share his views. If Musk wasn’t planning on leaving his post as a special government employee after the 130-day limit comes up, he might find a more persuasive business reason that it’s time to get back to his day job.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||