BUSINESS

Looking for help with today’s New York Times Wordle? Here are some expert hints, clues and commentary to help you solve today’s Wordle and sharpen your guessing game.

Scott Galloway, an NYU Stern professor and serial entrepreneur, used the June 8 Office Hours episode of The Prof G Pod, “How to Fix the Tax Code + the Problem With Corporate Jargon,” to lay out a detailed case against the current estate tax exemption. Galloway called it the centerpiece of a system that rewards inherited fortunes while forcing wage earners to pick up the tab. His argument arrives as the sheer scale of intergenerational wealth transfers draws increasing attention from economists, financial planners, and policymakers on both sides of the aisle.Galloway targets the $30 million estate exemption as a dynasty engineThe One Big Beautiful Bill Act, signed into law on July 4, 2025, permanently raised the federal estate and gift tax exemption to $15 million per individual, or $30 million for married couples starting in 2026.The law replaced a temporary provision under the 2017 Tax Cuts and Jobs Act that was set to expire at the end of 2025. Davis+Gilbert noted in its analysis that the sunset would have reduced the exemption to approximately $7 million per person in 2026.Galloway views that threshold as a policy failure, arguing it allows massive fortunes to pass across generations while ordinary workers face full taxation on every dollar they earn from labor.“We don’t need dynasties in the United States,” Galloway said on the podcast, calling the divide between American and European approaches to inherited wealth a direct threat to meritocracy.He proposed cutting the exemption from $30 million per couple to $1 million, a change he estimated would affect roughly 8% of households nationwide.The $84 trillion wealth transfer is fueling Galloway’s argumentCerulli Associates’ projected that $84.4 trillion in wealth would transfer through 2045, with $72.6 trillion flowing directly to heirs, and that 1.5% of American families would account for 42% of transfers.Cerulli has since revised that estimate upward, projecting $124 trillion in transfers through 2048 in its 2024 U.S. High-Net-Worth and Ultra-High-Net-Worth Markets report.Eventually, most of the wealth owned by older generations in the U.S. will be either donated or passed down to Gen X or Millennial heirs.Under the current estate tax structure, fewer than 1 in 1,000 estates are subject to federal estate tax, roughly 0.07% to 0.14% of decedents, according to Tax Policy Center and Congressional Research Service estimates, leaving the vast majority of this generational wealth effectively untouched by the system.

America’s $124 trillion inheritance wave is accelerating, with most family fortunes transferring untouched by estate taxes.Thomas Barwick/Getty Images

Galloway rejects the national sales tax as regressive and punishing for low-income householdsBefore presenting his own reform proposals, Galloway spent significant time on the podcast dismantling the case for a national sales tax, a concept that has gained periodic traction through the Fair Tax Act.He cited Bureau of Labor Statistics data showing that the lowest-income quintile reports average annual expenditures of $35,046, a level that exceeds their pre-tax income and forces many households to draw down savings or take on debt to cover the gap.A flat consumption tax applied to that spending would capture nearly all of their disposable resources while barely denting the budgets of the country’s highest earners, Galloway argued on the podcast.More Personal Finance:Fidelity has a warning for anyone who left a 401(k) at an old jobLiving trusts: what they do and who needs oneFidelity sounds alarm on 401(k)s, IRAs ACEEE research shows low-income households spend an average of 17.8% of their income on combined home energy and transportation fuel costs, and a consumption tax would add further pressure on housing, transportation, and utilities. “It’s not where we need to tax,” Galloway said. “We need people who are wealthy to feel it.”The Tax Foundation’s December 2023 assessment of the Fair Tax Act found that even with a monthly prebate intended to offset its regressive impact, the proposal would disproportionately burden retirees, large families, and students, Galloway explained.Galloway used Ronald Reagan to argue for equal capital gains ratesGalloway’s second proposal targets the long-standing gap between capital gains tax rates and ordinary income tax rates, which he framed as a structural advantage for people whose existing wealth generates additional wealth.He pointed to an unlikely historical precedent, referencing the Reagan era when capital gains and ordinary income were both taxed at the same rate under the federal code. He also took aim at Section 1202, which allows holders of qualifying small business stock to exclude up to $15 million in capital gains from federal taxation under the expanded rules in the One Big Beautiful Bill Act, Holland & Knight noted in a July 2025 client alert.The deficit math raises pressure beyond Galloway’s podcastThe fiscal implications of the expanded exemption are drawing scrutiny from nonpartisan research institutions tracking the long-term effects of federal tax legislation.The Yale Budget Lab estimated that the One Big Beautiful Bill Act would add $2.4 trillion to federal deficits as conventionally scored, with the debt-to-GDP ratio reaching 194% by 2054 under dynamic projections of the bill as enacted.In fiscal year 2024, federal estate and gift tax revenue combined totaled approximately $32 billion, according to the Bipartisan Policy Center, a figure that seems modest compared with a national debt that now exceeds $39 trillion.Galloway’s proposed estate tax overhaul would reach roughly 8% of households, a large expansion from the current 0.07%-0.14% range, while concentrating its impact on the wealthiest 2% of households, which Cerulli projects will account for roughly half of intergenerational asset flows through 2048.Related: When to buy a home instead of continuing to rent, according to Scott Galloway

On “Forbes Talks,” Forbes Executive Editor Luisa Kroll and Forbes Reporter Matt Durot discuss Elon Musk becoming the world’s first trillionaire after the IPO of SpaceX.

Moonshot AI released Kimi K2.7-Code this week, an open-source update to its K2 coding model family, claiming leaner reasoning and double-digit performance gains.K2.7-Code is built on the same trillion-parameter mixture-of-experts architecture as its predecessor K2.6, and drops in via an OpenAI-compatible API — which matters for teams already running K2.6 in production gateways.When K2.6 launched in April, it topped OpenRouter’s weekly LLM leaderboard — a ranking based on actual API routing decisions by developers, not self-reported benchmark scores.Moonshot AI says K2.7-Code addresses what it calls “overthinking,” reducing thinking-token usage by 30% compared to K2.6 — a number that would directly affect inference costs for teams running agentic workflows. Whether that efficiency gain holds on independent benchmarks is a question practitioners have already started raising publicly.What Kimi K2.7-Code isK2.7-Code is released under a Modified MIT license, with weights available on HuggingFace. The model is deployable via vLLM or SGLang. It runs exclusively in thinking mode and does not support temperature adjustment — Moonshot AI has fixed it at 1.0, meaning teams cannot tune output determinism the way they might with other models.The core change from K2.6 is how the model generates low-level code. Where K2.6 produced implementations by wrapping existing libraries and routing through established frameworks, K2.7-Code authors implementations directly. Moonshot AI says this produces more reliable generalization across Rust, Go and Python, and across task types including frontend development, DevOps and performance optimization.On benchmark performance, Moonshot AI claims gains of 21.8% on Kimi Code Bench v2, 11% on Program Bench and 31.5% on MLS Bench Lite. All three are proprietary benchmarks run by Moonshot AI. The model has not been submitted to DeepSWE, an independent coding benchmark that produces a 70-point spread across models — compared to SWE-Bench Pro’s 30-point spread — making it a more discriminating signal for teams configuring model routing systems.More honest, weaker for itThe picture from outside Moonshot’s own benchmarks is more complicated.Researcher Elliot Arledge ran K2.7-Code against K2.6 and Claude Fable 5 on KernelBench-Hard, a public benchmark focused on GPU kernel optimization, and published his full run logs at kernelbench.com. “K2.7 is more honest but not more capable,” Arledge wrote on X. On five of six problems, K2.7-Code produced real authored Triton kernels where K2.6 had used library wrappers. Two of those kernels failed on the model’s own bugs. The MoE kernel result regressed from K2.6’s score of 0.222 to 0.157. “Fable, for reference, tops every cell it doesn’t honestly fail,” Arledge wrote.Sugumaran Balasubramaniyan, a developer who built a model-task-router for the Hermes Agent platform using DeepSWE as his reference signal, responded publicly to the K2.7-Code release and challenged Moonshot AI directly on the benchmark choices. “Respectfully, every model ‘improves’ double digits on its own test suite,” Balasubramaniyan wrote on X. He noted that K2.6 scored 24% on DeepSWE, tied with GPT-5.4-mini, and asked whether Moonshot AI would submit K2.7-Code to the same benchmark. Balasubramaniyan said it took 13 review rounds to get the benchmark data right for his router and that he would route coding tasks to K2.7-Code if the independent numbers hold up.What this means for enterprisesThe token efficiency gain is immediately usable. Teams running K2.6 in production can swap in K2.7-Code via the OpenAI-compatible API and expect lower inference costs on agentic workflows without an architecture change. The 30% thinking-token reduction is Moonshot’s own number, but the integration path is low-risk enough to test against your own workloads before committing.The practical question is whether those efficiency gains hold on a team’s own task distribution. Running K2.7-Code against your own workloads before adjusting gateway weights is the low-risk path to finding out.

Large language models continue to struggle with hallucinations, presenting a major roadblock for real-world enterprise applications. Reducing these errors is a messy business, forcing model developers to navigate a strict tradeoff where eliminating factual errors often suppresses valid answers.In a new paper, Google researchers introduce the concept of “faithful uncertainty,” a metacognitive technique that aligns a model’s response with its internal confidence. This alignment allows the model to offer appropriately hedged hypotheses, such as “My best guess is,” instead of defaulting to an unhelpful “answer-or-abstain” binary.In real-world agentic AI applications, this metacognitive awareness acts as an essential control layer. It empowers autonomous systems to accurately determine when their internal knowledge is sufficient and when they must dynamically trigger external tools or search APIs to resolve deficits.The utility tax of current mitigation strategiesUnderstanding why LLMs hallucinate hinges on separating two capabilities: a model knowing facts versus knowing what is known. Historically, most factuality gains in AI have come from expanding the knowledge boundary, meaning developers simply pack more facts into the model’s parameters through larger scale and more training data.However, expanding a model’s knowledge does not automatically improve its boundary awareness, which is its ability to distinguish the known from the unknown and recognize its own limitations.“There are broadly two ways to improve LLM factuality,” Gal Yona, Research Scientist at Google and co-author of the paper, told VentureBeat. The first is continuing to teach the model more facts. But, Yona notes, “model capacity is finite, and the long tail of knowledge is effectively infinite.” Once models hit this limit, the hope is they know what they don’t know and simply abstain from answering. However, this is inherently difficult for LLMs.“This is why most practical attempts to reduce hallucinations through various interventions don’t actually make it to deployment,” Yona explains. “They do reduce hallucinations, but they also hurt utility, because the model ends up refusing to answer questions it actually does know.”This inability to distinguish between knowns and unknowns creates what the paper’s authors call the “utility tax.” Enforcing a zero-hallucination standard requires the model to abstain whenever it is even slightly uncertain, discarding massive volumes of completely valid information. For example, the authors demonstrate that reducing an underlying 25% error rate down to a strict 5% target forces developers to discard 52% of the model’s correct answers.Treating all errors as hallucinations forces enterprise systems to choose between trustworthiness and helpfulness. Application developers are generally unwilling to pay this massive utility tax and render their models unhelpful. Consequently, they optimize systems to prioritize coverage, forcing models to operate in a state where they continue to generate confident hallucinations.Reframing hallucinations as confident errorsTo move past the utility tax, the researchers propose to stop treating any factual error as a hallucination. Instead, they reframe hallucinations as “confident errors”: incorrect information delivered authoritatively without appropriate qualification.This subtle reframing dissolves the strict “answer-or-abstain” dichotomy and allows the model to express its uncertainty. In this new framework, if a model makes a factual mistake but appropriately hedges its response (e.g., by stating, “I am not completely sure, but I think…”), it isn’t a hallucination. It is simply a hypothesis offered to the user for consideration. By expressing uncertainty, the AI preserves its utility—sharing whatever partial or likely knowledge it has—without violating the user’s trust.However, if an AI assistant hedges all its responses with a disclaimer, the user is forced to double-check everything, defeating the purpose of the tool entirely.The solution the researchers propose is “faithful uncertainty.” This approach requires aligning a model’s linguistic uncertainty, or the words it uses to express doubt, with its intrinsic uncertainty, which is its actual, internal statistical confidence in that specific answer. This ensures the model only hedges when its internal state genuinely reflects conflicting or low-probability information.Faithful uncertainty forms a core component of “metacognition,” the AI’s ability to be aware of its own uncertainty and act on it. To understand this practically, consider the intuitive example of consulting a doctor. We do not trust doctors because they are all-knowing. We trust them because they reliably distinguish between a confident diagnosis (“You have a fracture”) and an educated hypothesis (“It might be a sprain, but let’s run some tests”).Practical implications for enterprise AIUnder the new framing, errors where a model is genuinely confident but factually incorrect are categorized as “honest mistakes.” This casts knowledge expansion (training the model on more data) and faithful uncertainty as completely complementary efforts. Knowledge expansion pushes the absolute knowledge boundary outward to minimize honest mistakes, while faithful uncertainty honestly communicates wherever that boundary currently lies.This new framing has important implications for agentic applications. The shift to agentic AI might make it seem like knowing what the model doesn’t know is redundant, since models can just search external databases. However, access to external tools actually amplifies the need for faithful uncertainty. In agentic systems, metacognition becomes the central control layer that governs the entire system.External tools solve the storage problem because the model no longer needs to encode every fact into its parameters. However, this introduces a new control problem: managing when to retrieve information, verify facts, and orchestrate these external tools. Without faithful uncertainty, an agent is essentially flying blind and must rely on external, static heuristics or over-engineered scaffolds.“The model might search for something it already knows confidently—wasting latency and cost for no gain. Or the opposite: it confidently answers from memory when it should have searched, producing a plausible but wrong output,” Yona said. Today’s agent harnesses try to solve this externally with query classifiers or always-search rules, but Yona notes that these are “static and brittle.” By using its intrinsic uncertainty to regulate its own behavior, the agent dynamically optimizes its tool use, choosing to invoke a search tool only when its internal confidence is genuinely low.Beyond deciding when to search, faithful uncertainty is critical for evaluating the results of a search. If a tool returns low-quality or unexpected information, a metacognitive agent does not blindly accept whatever appears in its context window. Instead, it uses its uncertainty awareness to weigh the retrieved external signals against its own internal priors. This prevents sycophantic behavior where the system might otherwise trust external sources that conflict with its actual known knowledge.The bootstrapping paradox: The catch to teaching uncertaintyFor enterprise builders, achieving this faithful uncertainty is trickier than it sounds. It requires teaching models the syntax of uncertainty through supervised fine-tuning (SFT). Because pre-trained models are mostly fed authoritative text, they must be explicitly taught to say things like, “I’m not entirely sure, but I think VentureBeat was founded in…”But SFT introduces a “bootstrapping paradox.” Unlike standard training datasets where the “right answer” is the same regardless of the model, the ground truth for uncertainty is the model’s own dynamic knowledge base.“Here’s the catch: the ‘correct’ expression of uncertainty is inherently dynamic, because it depends on what this particular model knows or doesn’t know at this particular point in training,” Yona said. “If you train on a label that says ‘I don’t know X’ but the model actually does know X, you’ve taught it to hallucinate uncertainty… The training data is static, but the target is a moving one, and that’s the fundamental tension teams need to grapple with.”The road to self-aware AIFor enterprises looking to implement these capabilities without expensive retraining, prompting serves as the most accessible entry point. “Prompt engineering is already something most engineers do today, this provides the lowest-friction path to improving metacognitive behavior today,” Yona said. Enterprise developers can explore frameworks like MetaFaith, an open-source project previously co-authored by Yona, to begin applying metacognitive prompting to off-the-shelf models.However, Yona cautions that “there is still substantial headroom that prompting alone doesn’t solve,” meaning the industry will eventually need to rely on advanced reinforcement learning (RL) to bake metacognition deeply into model training.Ultimately, as enterprises transition from isolated chat applications to complex, multi-agent workflows, self-awareness will become a defining prerequisite for reliable autonomy. But evaluating whether a model truly possesses this awareness remains a profound technical challenge.“How do you actually evaluate whether a model can sense its internal states?” Yona asks. “Even in humans, it’s hard to define or separate ‘true’ self-monitoring abilities from a capable reliance on proxies. We face exactly the same challenges with LLMs: a model might learn to mimic the style of uncertainty without truly sensing its internal state. Developing evaluation frameworks that can tell the difference is one of the most important open problems in this space.”

There were a lot of ways that SpaceX’s initial public offering could have gone wrong. Instead, the company bucked Wall Street norms, pulled off the biggest IPO ever and raised $75 billion.

After years of expansion, few would have predicted that the fitness chain that helped popularize boutique indoor cycling would now be facing another round of studio closures across the U.S.Founded in 2006 in New York City, SoulCycle grew from a single studio with a small number of bikes into one of the most recognizable fitness brands in the industry.Its signature 45-minute rhythm-based indoor cycling classes, set to curated music playlists and built around instructor-led experiences, attracted a devoted customer base and helped expand the company to more than 60 locations across the U.S., Canada, and the U.K.Now, SoulCycle appears to be continuing a yearslong effort to reduce its physical footprint.SoulCycle is closing multiple studios nationwideSeveral SoulCycle locations are scheduled to close on June 14, according to local reporting and employee communications reviewed by TheStreet.The affected studios include:Walnut Creek: 116 Broadway Lane, Walnut Creek, California (reported by Walnut Creek Spotlight)La Jolla: 4303 La Jolla Village Dr, Suite 2110, San Diego, California (based on an employee post)Denver: 265 St. Paul St, Denver, Colorado (based on an employee post)Bryant Park: 110 W 41st St, New York, New York (based on an employee post)Manhattan Beach: 820 Pacific Coast Hwy #106, El Segundo, California (based on an employee post)South Beach: 2325 Collins Ave, Miami Beach, Florida (based on an employee post)SoulCycle has not publicly announced the closures. However, employees at several affected studios have reportedly received notices regarding the shutdowns. Additional closures have circulated in online discussions, although no further studio shutdowns have been officially confirmed at the time of publication.TheStreet contacted SoulCycle for comment and will update this story if a response is received.Why is SoulCycle closing locations?The latest closures come during a period of leadership transition and continued pressure across the fitness industry.Earlier this month, SoulCycle announced that CEO Evelyn Webster would step down after nearly six years to become chief executive of podcast company Audiochuck, according to a company press release.The leadership change follows several years of broader restructuring across the business.Like many fitness companies, SoulCycle faced major disruptions during the Covid pandemic, when many operators temporarily suspended in-person classes due to stay-at-home orders, prompting a rapid shift in consumer workout habits.Its parent company, Equinox Group, reportedly recorded approximately $350 million in losses during 2020 tied to pandemic-related disruptions, Bloomberg reported.SoulCycle has also reduced its physical footprint in recent years. In 2022, the company closed 20 studios, roughly one-quarter of its locations at the time, as part of broader operational changes, reported CNN.While SoulCycle has not publicly linked the current closures to any specific reason, the move aligns with an industry-wide push to balance operating costs with changing consumer demand.

SoulCycle is closing locations nationwide.Stephen Lovekin/Getty Images

The fitness business continues to evolveSoulCycle’s challenges reflect broader shifts across the fitness market rather than changes isolated to a single company.Peloton (PTON), one of SoulCycle’s best-known competitors in cycling-based fitness, built its business around connected equipment and instructor-led classes delivered at home rather than in physical studios.The company experienced rapid growth during the pandemic as demand for at-home workouts and digital fitness subscriptions surged.But as restrictions eased and consumers returned to gyms and in-person experiences, growth slowed significantly.Here’s some of my previous coverage on closures:Iconic gym chain shuts down 23 locationsConvenience store giant sells stores, exits market102-year-old fashion giant faces 400 store closuresSince 2021, Peloton has focused heavily on restructuring efforts. In fiscal 2025, the company reduced overall expenses by 25%, cutting spending across sales, marketing, research and development, and administrative operations. It also reduced its retail footprint, closing 24 of 37 showrooms.In 2026, Peloton announced additional cost-cutting measures, including workforce reductions affecting approximately 11% of employees globally, Reuters reported.Although Peloton recorded a 1% year-over-year increase in total revenue during the third quarter of fiscal 2026, membership trends remained under pressure, with total members declining 5% and paid app subscriptions falling 9%.Industry participation data suggests consumer preferences may also be changing.According to a 2025 report from the Sports and Fitness Industry Association (SFIA), pilates saw one of the fastest growth rates among fitness activities over the previous five years, increasing nearly 40% since 2019.Meanwhile, cycling participation experienced one of the largest declines among tracked activities, down 33.5% during the same period.That does not necessarily mean cycling fitness is disappearing. But the data suggest that fitness operators across both studios and at-home categories are adapting to a market where consumer habits look different from those during the previous decade of growth.Related: After bankruptcy, iconic seafood chain closes flagship restaurant

Thanks to a landmark 2022 agreement, U.S. women’s soccer players will get part of the World Cup prize pool, too. Here’s how it works.

The Lincoln Center for the Performing Arts honored Misty Copeland with the inaugural Luminary award at their Summer Gala

Co-founding Soundgarden guitarist Kim Thayil on his new book ‘A Screaming Life,’ Soundgarden’s DIY approach to business and completing the final Soundgarden album.