AT&T

T-Mobile Cut a Cable to Expel Chinese Hackers

TechCrunch reports, citing Bloomberg, that T-Mobile identified and expelled Chinese government-backed hackers from its network in 2024 by physically severing a network cable to isolate a compromised system. The intrusions were part of a wider Salt Typhoon campaign that targeted hundreds of telecoms, internet companies and datacenter providers to collect phone records and information on senior U.S. officials. T-Mobile’s cybersecurity chief Jeff Simon and colleagues reportedly located the compromised box at a Bellevue, Washington data center and cut the external connection. Other affected companies cited include AT&T, Verizon, Viasat, Charter and Windstream.

What If California Blocks Paramount–WBD Deal?

If a California-led state antitrust challenge succeeds in blocking Paramount Skydance's $110.9 billion acquisition of Warner Bros. Discovery, WBD is widely expected to pursue a major restructuring. Likely outcomes include spinning off the Global Linear Networks as a separate publicly traded company, selling the studio and streaming assets to tech-first streamers (e.g., Netflix, Amazon, Apple), or executing piecemeal asset sales (with Comcast previously expressing interest in studios). The article notes the transaction cleared federal review and shareholder approval but faces a March 2027 state trial; any successful state action would force WBD’s board to revisit alternatives, reallocate debt between split entities, and could reshape the streaming/linear landscape over 12–18 months.

Economics of Open vs Closed AI Models

An opinion analysis by Christian Catalini (published on a16z on 2026-08-11) examines the economic trade-offs between open-weights and closed AI models. Using historical analogies (Great Exhibition, Celera vs Human Genome Project, AT&T), the piece argues that open weights are unlikely to change the overall level of AI investment but will redirect where innovation occurs, who builds applications, and who captures returns. The article frames safety concerns (e.g., Anthropic and Dario Amodei’s warnings about distillation and misuse) alongside arguments that openness can accelerate diffusion, broaden defensive capabilities, and reduce vendor lock-in for enterprises. Complementary assets (data, distribution, tacit knowledge) and regulatory capture are highlighted as decisive factors determining whether value accrues to frontier labs or to downstream integrators. Concludes both open and closed approaches may coexist, with important domain-specific differences for security and commercialization.