 |
|
|
|||||
Chatbots and AI copilots that can write fluent emails and computer code exploit statistical probabilities in natural or artificial language to predict the next token in any given sequence. But the moment you try to breach that boundary of language, you start to run into problems, and that’s what people are running into now, says Rohan Kodialam, a former data scientist at Citadel. If you try to use [generative AI] on a real business application where it’s not just about saying stuff but about doing stuff, all of a sudden you are now in a much more difficult spot. The problem helps explain growing skepticism about the global AI boom. Last month an MIT project on the agentic web reported that despite up to $40 billion in enterprise investment, 95% of organizations are seeing no return. The studys evidence was thin, but U.S. Census Bureau data does show fewer large companies adopting AI in their operationsthe first sustained decline since ChatGPTs 2022 debut. To address that gap, Kodialam and engineer Jamie Bloxham cofounded Sphinx, a startup building AI designed to work reliably with data. The seven-person firm, based in Queens, New York, last week raised a $9.5 million seed round led by Lightspeed, with backing from Bessemer Venture Partners, BoxGroup, K5, Impatient Ventures, Steve Cohen, Naveen Rao, and others. Kodialam is Sphinxs CEO; Bloxham is CTO. Designed for data scientists, Sphinxs copilot targets a surprising gap in the LLM market. Over the past year, tools like Cursor and Windsurf have reshaped software engineering workflows. But Kodialam says that even in quantitative fields where machine learning is standard, data specialists lack the kinds of AI copilots available to others. Its super frustrating to see that all the software engineers have Claude Code, all the front-office guys have ChatGPT,” he says, “and the data people have nothing, because no one has built AI that interfaces with data. To appreciate the depth of AIs shortcomings, consider a problem Kodialam calls the gorilla in the data. As part of a 2020 study, New York University professors gave data to students and claimed it showed a population’s steps walked versus body mass index. In reality, when the data was plotted on a graph, it resembled a picture of a gorilla. Only some students noticed the joke; others reported a negative correlation. “The professor told them, You obviously should have looked at the data and seen that it was a monkey and concluded that this was useless data, Kodialam says. When researchers recently gave the same problem to LLMs like GPT-4o, many models were similarly duped. Addressing the gorilla in the data means building the bridge between the modality of text and images, which is what LLMs live in, and then the modality of numerical tabular data, which is where data lives. Humans and LLMs can miss the gorilla hiding in the data. [Photo: Courtesy of Sphinx] Sphinxs AI copilot integrates with Jupyter Notebook, the open-source environment widely used for exploring datasets. It can clean and analyze data, build visualizations and models, and, in a “fully agentic loop,” break tasks into steps executed cell by cell while generating and reading charts to guide its progress. In July, the framework outperformed both general-purpose frontier models and domain-specific tools like Googles Gemini Data Science Agent on DABstep, a benchmark for multistep data reasoning. With Sphinx, you just say, Okay, segment the customers in a reasonable way. It does it, Kodialam says. You say, Actually, I don’t like that clustering. Can you make it better in this way or that way? That speed and ease of use has already shown results in practice. Brian Tate, CEO of drinkable oatmeal brand Oats Overnight, said in a statement that Sphinx has helped his companys data scientists uncover patterns in shopper behavior in minutes instead of hours or days. Sphinx isnt the only firm trying to use foundation models and tools like vector databases to help customers wrangle sprawling and disparate datasets. And demand is rising, as executives tryfor real this timeto make foundation models actually worthwhile. Business intelligence vendors like Tableau and Palantir offer AI tools and elaborate dashboards; more technical data vendors, like Snowflake and Databricks, have been adding features to allow nonexperts to analyze data. (Sphinxs investors also include executives at Databricks, Windsurf, and Together AI.) The AI lessons Sphinx is building on Dashboards can be nice to look at, but Kodialam sees more value in asking the right question at the right time and getting an answer very quickly. You talk to leaders, he says, and they’re like, Yeah, I look at like about 2 of the 500 dashboards in my company ever. Most of the time a dashboard is built, I look at it once and then I move on, because the reason the dashboard is built is to answer an acute question. That’s obviously not what a dashboard is for, and answering acute questions is what Sphinx is trained at. How else does Sphinx differ from a stand-alone system like Palantirs? Kodialam emphasizes Sphinxs quick onboarding process, and its fluid, semantic approach to dynamic datasets. You shouldnt need to have a complicated, human-driven, brittle process to organize your data, he says. The AI should organize your data on its own. Kodialam says hes bringing a few other big lessons with him from Wall Street. One: Complex AI-based processes are both powerful and fragile. Small changes to the system as a whole can lead to massive performance increases, he says. But this is a double-edged sword. If you get it right, you can do great things. But also constant vigilance is needed, and every single part of the stack can break the entire process. The second lesson is that data itself is the real advantage. Companies with vast reserves will lways have the upper hand, while newcomers face an uphill climb. As Kodialam puts it, Incumbents with a lot of data are going to do a really great job at deploying AI, and people who are just kind of emerging as startups without a lot of data are probably going to have a hard time. Relatedly, Kodialam stresses the importance of privacy. Data, he says, is often viewed as a companys most valuable asset, and firms will only trust an AI partner if their information is handled like a protected resource rather than something stored or reused after passing through an API. One basic challengeand opportunityis that many companies with plenty of data dont examine it closely. Kodialam says data scientists should “spend their time thinking about, What is a hypothesis that I need to think about for my business? What is the point of this hypothesis? How does it make dollars? How does it improve the bottom line? How do I interpret it? He adds: They can focus on that layer, where their intelligence and creativity and understanding of their business actually adds value, and get things done incredibly quickly once they have those ideas.” One big challenge for Sphinx now is incorporating the data thats not yet data: the tribal knowledge in an institution, the know-how of some person who’s been there for 10 years and just happens to know the secret way to interpret the data that will make it sing. But that data is often in peoples heads, not written down. Like a junior hire, Sphinxs copilot can be trained to incorporate this process knowledge into its agentic framework too. There is not really a lot of data on data science, Kodialam says, which is kind of ironic, given that it’s data about data. Kodialam honed his pragmatic approach to AI while working with healthcare data as a student at the Massachusetts Institute of Technology and, later, as a quantitative researcher at Citadel, Wall Street’s most profitable hedge fund, using torrents of structured and unstructured data to beat the market. He ended up running a research team focused on AI, where there was little room for error or hallucination. With a lot of AI applications right now, you can just generate something cool and that’s good enough. But in finance, you will not make money if you’re wrong,” he says. Like the models, the humans who keep those models up-to-date must be high-performing and willing to work long hours. No wonder, then, that big AI companies have been luring Wall Street quants with gigantic pay packages, part of an industry-wide talent blitz. Kodialam says hes lost candidates to AI giant Anthropic, but he gets it. The only real moat you have in AI is talented people, he says. Still, while many AI researchers may be focused on AGI, Sphinx hopes to attract candidates who want to do something useful now. Many of them come from industries like finance, where the reward function is a lot more short term, Kodialam says. You do the right thing and the alpha makes money tomorrow. As for the impact of Sphinxs AI tools on quants and other data professionals, Kodialam is optimistic. We see this as a way to make people more productive, not to replace anyone, he says. There is so much value to be had in most companies data if they’re able to optimize what they’re doingeverything from supply chain to pricing to marketingthat the existing cohort of data scientists could easily grow by 10x and still be delivering massive value. Kodialam’s obsession with the here and now comes with a longer-term vision about the future of AI. If foundation models cant grok data in all of its formsif they cant do the work of data sciencethey wont reach the next level of intelligence. The place where humans are really good compared to AI right now is that humans are very multimodal, he says. Naturally, if I give you a new modality, you can usually mess around with it and figure it out: audio, music, structured data, unstructured data, touch, smell. There are so many modalities that humans can just generalize over very quickly. AI is kind of stuck in whatever modality the person who trained it was able to tokenize. In the near term, Kodialam predicts that those kinds of shortcomings will make it easier to separate vaporware from the good stuff. A good AI system is very useful; a bad AI system is complete trash, he says. And there’s no difference in a lot of people’s minds because it’s just AI. But the quality matters so much. And I think that’ll just become increasingly obvious as AI gets deployed in every company, and some companies are not winning and others are.
Category:
E-Commerce
Samsonite, the world’s largest luggage brand, has just released a new suitcase design that’s an amalgamation of its best ideas from over its 115-year history. Named Paralux, it is lightweight yet hardshell, and its shell is made with 50% recycled polypropylene. “I want to think of it as our greatest hits album,” VP of design Ulliyada Bopanna told Fast Company. But if Paralux is Samsonite’s greatest hits album, it also has a bonus track. The collection features something that neither Samsonite nor any other brand has ever experimented with before: The hardshell suitcase sports two openingsone in the middle (like most standard suitcases), and one at the front (like some soft-shell carry-ons). The result is a dual-access suitcase that lets you access the full contents of your luggage by simply unzipping the front door in a cramped hotel room or when you’re navigating crowded airport terminals. [Photo: Courtesy of Samsonite] A suitcase with global appeal Paralux is the product of a rare collaboration between Samsonite’s various design teams across New York and Asia. The collaboration is unusual because, as Bopanna explains, luggage needs vary drastically from one region to another. In Europe, for example, travelers prioritize lightweight designs, largely because of the strict weight restrictions associated with low-cost airlines such as Ryanair or Wizz Air. In the U.S., “it’s the exact opposite,” says Bopanna, adding that American travelers perceive lightweight luggage as weak and fragile. These regional variances have meant that Samsonite often designs suitcases to meet the specific criteria of particular markets. But Paralux was intended as a global collectionone that could appeal to travelers across the world, especially those who have experienced, firsthand, the cramped quarters of hotel rooms in Asia. The collection includes a dual-access carry-on (starting at $319), two spinners in medium ($369) and large ($409) sizes, plus a duffel ($169), and a versatile two-in-one backpack that lets you pack clothes in the main compartment, then unzip a separate day bag for quick trips ($199). [Photo: Courtesy of Samsonite] Engineering the impossible Samsonite isn’t operating in a vacuum. Over the past decade or so, the luggage world has gotten crowded thanks to a parade of Instagram-ready startups that have turned suitcases into lifestyle statements. Away burst onto the scene in 2015 with sleek hard shells and built-in phone chargers, making $275 carry-ons feel aspirational. Then came Monos with its minimalist aesthetics, and Dagne Dover, which sought to make soft-side suitcases sexier. For a 114-year-old giant like Samsonite, standing out in this design-obsessed market isn’t just smartit’s survival. The brand, which commands nearly 20% of the luggage market with its brands including Samsonite, Tumi, and American Tourister, has seen its sales shrink in 2025. Net sales for the first half of the year are down 5.2%. The Paralux suitcase offers a moment of renewed innovation. The suitcase recently won two Red Dot Awards for sustainability design and overall product design. It took three years to develop, and more than 20 prototypes to get right. “There’s a good reason why people aren’t doing it,” Bopanna says with a laugh. “It’s because it’s bloody tough.” [Photo: Courtesy of Samsonite] Unsurprisingly, the dual-access feature was the biggest challenge to solve, since any incisions in a hardshell compromise the structural integrity of the shell. “Adding a front door creates weakness,” says Bopanna. “For carry-ons, that’s not such a big deal, but once you look at large luggage, it starts to torque easily.” The designers couldn’t just make some parts of the suitcase thicker, because that would have made it heavier, eliminating some of the appeal for weight-conscious travelers. At 11.2 pounds, the large suitcase is among the industry’s lightest, though not the absolute lightest. The Samsonite team is aware of this trade-off as the original weight goal was even lower, but they ultimately compromised on weight in favor of durability. Bopanna declined to share specific engineering details, but he noted that the entire design is meant to be self-reinforcing, including the parallel ridges on the front of the hardshell, which provide structural support. “If we’d chosen more robust materials or aluminum [which is heavier than polypropylene], it wouldn’t have been as difficult, but we made a conscious choice to take a step forard using recycled polypropylene,” he says. Since virgin polypropylene is slightly stronger than its recycled counterpart, the team used a higher ratio of recycled material in the shell, and chose entirely virgin poly for the wheels, which take the most beating. The handles, meanwhile, are made of recycled aluminum. The team also made the handles and wheels easier to repair by reducing the number of screws; if one of them breaks, they can ship you a spare part that you can install yourself. In the end, the Paralux collection exemplifies the brand’s broader approach to sustainability, which Bopanna describes as finding simple solutions to big problems. “Innovation doesn’t have to be big,” he says, “as long as you are solving a recurring problem, that to me is a great way to move the benchmark forward.”
Category:
E-Commerce
With the Federal Reserve widely expected Wednesday to reduce its key interest rate by a quarter-point to about 4.1%, economists and Wall Street investors will be looking for signals about next steps: How deeply might the Fed cut in the next few months?There are typically two different approaches the central bank takes to lowering borrowing costs: Either a measured pace that reflects a modest adjustment to its key rate, or a much more rapid set of cuts as the economy deteriorates in an often-doomed effort to stave off recession.For now, most economists expect the Fed will take the first approach: What many analysts call a “recalibration” of rates to keep the economy growing and businesses hiring. Under this view, the Fed would reduce rates as many as five times by the middle of next year, bringing its rate closer to a level that neither stimulates or slows the economy.Wall Street traders expect three reductions this year and then two more by next June, according to futures pricing tracked by CME Fedwatch.A rate cut Wednesday would be the first in nine months. The Fed, led by Chair Jerome Powell, reduced borrowing costs three times last year. But it then put any further cuts on hold to evaluate the impact of President Donald Trump’s sweeping tariffs on the economy.As recently as their last meeting in late July, Powell described the job market as “solid” and kept rates unchanged as officials sought to take more time to see how the economy evolved.Since then, however, the government has reported a sharp slowdown in hiring, and previous government data has been revised much lower. Employers actually cut back slightly on their payrolls in June, shedding 13,000 jobs, and added just 22,000 in August.The government also said last week that its estimate of job gains for the year ended in March 2025 would likely be revised down by 911,000, a sharp reduction in total employment. Powell and other Fed officials had previously pointed to a robust job market as a key reason that they could afford to keep rates unchanged. But with businesses pulling back on hiring, the economic case for a rate cutwhich can spur more borrowing and spendingis stronger.The downward revision of nearly a million jobs is a “huge downgrade,” said Talley Leger, chief market strategist at the Wealth Consulting Group. “If that doesn’t light a fire under the Fed just from an economic perspective I don’t know what will.”Still, inflation remains stubbornly elevated, partly because tariffs have lifted the cost of some goods, such as furniture, appliances, and food. Prices rose 2.9% in August from a year earlier, the government said last week, up from 2.7% a month earlier.Persistent inflation could keep the Fed from cutting too rapidly. The central bank will release its quarterly economic projections after the meeting Wednesday, and many economists forecast they will show that officials expect three total reductions this year and at least two more next year.Five reductions would bring the Fed’s key rate down to just above 3%. Many economists think that is roughly the rate that would neither stimulate nor slow the economy.If Fed officials began to worry the economy would slip into recession, they would likely cut rates more quickly. But for now, most economists don’t see rapid cuts as necessary.“We’re not at a break-glass moment,” said Vincent Reinhart, chief economist at BNY Investments. “This is a recalibration.” Christopher Rugaber, AP Economics Writer
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||