Coinbase Empowers AI Trades, Google DeepMind’s Tactical AI Revolution, Anthropic's Controversial Claude Corps, Amazon Axes AI Leaderboard Scandal

**SPEAKER_2** (0:05)
Welcome to Digimasters Shorts.

**SPEAKER_3** (0:07)
We are your hosts, Adam Nagus and Carly Wilson, delivering the latest scoop from the digital realm.
Coinbase has launched a new product called Coinbase for Agents, designed to let artificial intelligence systems connect directly to users' Coinbase accounts and execute financial transactions on their behalf. The platform enables AI agents such as ChatGPT and Claude to trade cryptocurrencies, access market data, pay for online services and eventually make purchases within user-defined risk and spending limits. Coinbase says the system allows users to issue natural language commands, authorizing tasks like portfolio rebalancing and automated trading strategies. At launch, agents can trade spot crypto and derivatives, with plans to expand into equities and prediction markets. The company is also integrating support for X402, a machine-to-machine payments protocol that allows agents to make small payments for services like premium data and computing without subscriptions. Coinbase describes the move as a step toward agentic commerce where AI systems handle a growing share of financial activity. The firm points to forecasts suggesting autonomous agents could account for up to 20 percent of e-commerce activity by 2030
To address security concerns, agents can operate within isolated portfolios and will soon include customizable controls such as spending caps and service restrictions. The launch comes as Robinhood recently introduced a similar product and as tokenized assets and stable coins continue to reach record market capitalizations.

**SPEAKER_2** (1:34)
Google DeepMind has developed an artificial intelligence system called TacticAI that can predict football plays up to eight seconds before they unfold. Built using geometric deep learning, the system models player movement from standard broadcast footage and recommends tactical adjustments in real time. It was originally created in collaboration with Liverpool FC where club experts preferred the AI's tactical suggestions over real match setups 90% of the time. The research published in Nature Communications also showed the system outperformed existing models in predicting corner kick outcomes and shot likelihood. Now, Brazilian club Palmeiras has become the first team to use TacticAI for live open play analysis rather than just set pieces. Coaches can reposition players using a drag and drop interface and instantly see how those changes affect both teams' overall structure.
The technology allows staff to test scenarios such as pushing a defender higher up the pitch and measure the tactical ripple effects across the field. Google has also partnered with Brazil's Football Confederation to explore AI applications ahead of the World Cup.
Experts say the breakthrough could signal a broader shift as football begins to adopt advanced AI-driven tactics similar to systems already used in baseball and basketball.
Anthropic has announced a $150 million initiative called Claude Core to place 1,000 early career AI fellows inside more than 400 nonprofit organizations across the United States. The fellows will earn $85,000 plus benefits for a year-long placement focused on helping nonprofits integrate and use Anthropic's AI model Claude. No college degree is required, but applicants must be at least 18 years old, have US work authorization, and possess no more than two years of full-time work experience. Applications closed July 17th with the first cohort of 100 fellows scheduled to begin in October 2026, and additional groups launching in 2027 Each participating nonprofit will also receive a $10,000 grant and free Claude credits with recruitment and training managed in partnership with CodePath. Anthropic says the program is designed to expand access to AI tools for organizations that typically lack resources and technical expertise to adopt them. The initiative mirrors service programs like AmeriCores, but places a corporate AI product at the center of its mission. Critics argue the effort doubles as a distribution strategy, training nonprofits to build workflows around Claude while positioning the company as a philanthropic leader. The launch comes as Anthropic prepares for a potential IPO and seeks to distinguish itself as a responsible AI provider. Whether Claude Core becomes a lasting force for nonprofit innovation or primarily a strategic branding move may depend on its long-term impact beyond the first 1000 fellows.

**SPEAKER_3** (4:20)
Amazon's decision to dismantle an internal AI leaderboard has spotlighted a growing problem in corporate America's approach to artificial intelligence. The tracker measured so-called token maxing where employees maximized their AI usage to signal productivity, but many were simply gaming the system. Industry experts say the episode reflects a broader mistake, treating AI as a speed tool layered on to outdated processes rather than a catalyst for true business transformation.
At the Fortune Brainstorm Tech Conference in Aspen, leaders argued that companies are over-investing in technology while under-investing in people. Deloitte's China Widener said only 7 cents of every AI dollar goes toward human support while the rest funds the technology itself. That imbalance has concentrated AI's economic rewards among a small group of resource-rich firms, with a PWC study finding just 20% of companies capture nearly three-quarters of the value. ServiceNow's Chris Bedi noted that about 90% of enterprise AI efforts focus on cost-cutting and internal productivity instead of revenue growth. He and others urged executives to move beyond tracking hours saved and instead pursue measurable organisational outcomes. Former Paramount Technology Chief Phil Weiser recommended deploying centralised engineering teams inside business units to fundamentally rethink operations. Experts added that scepticism among younger workers and resistance from seasoned professionals may be the biggest barrier to AI adoption.