Daily English Article — October 28, 2026

Tesla's long-delayed second-generation Roadster generated fresh excitement in mid-October when Elon Musk demonstrated the car briefly becoming airborne using a compressed-air thrust system. The flight lasted only a few seconds, and recharging the system takes six to seven hours. Despite these limitations, some buyers immediately moved to secure their reservations, captivated by the spectacle. Alex Lawrence, who owns an electric vehicle dealership called EVAuto, watched the live demonstration and placed a second reservation on the spot.

The Roadster has had a notoriously turbulent path toward production. Tesla first unveiled the second-generation model as a surprise at a 2017 event and began accepting $50,000 deposits immediately, with an initial target delivery date of 2020. That timeline has shifted dramatically, with Tesla now projecting production to begin in 2028 and delivery dates remaining unclear. To put the opportunity cost in stark perspective, reservation holders who had invested that $50,000 in Tesla stock instead would now be sitting on nearly $1 million.

The Roadster lineup is divided into multiple tiers with vastly different price points. The base model requires a $50,000 reservation deposit and excludes the thruster system entirely. The mid-tier Thrust variant, which includes the compressed-air acceleration system, demands a $250,000 deposit. The top-tier Rocket model—featuring the full flight capability—is listed as invite-only. Musk himself has acknowledged that this version will not be street legal, and purchase prices across all variants are described as estimates, with no confirmed figures yet published.

Lawrence sees the Roadster primarily as a business asset. His EV dealership boasts a large and rapidly growing social media following, and the reservations were technically placed in the company's name. He plans to use the vehicles for marketing content and track demonstrations rather than daily transportation. He remains determined to acquire the flying Rocket variant in particular, viewing it as a powerful tool for generating online engagement and brand visibility, regardless of the considerable financial and regulatory uncertainties involved.

Travis Cook, a Utah-based tech founder and self-described car aficionado, upgraded his nearly decade-old base reservation to the Thrust tier because he genuinely wants to drive the car. Holding a dealer's license, Cook regularly buys and sells unusual vehicles, and he relishes the social reactions they attract. He acknowledges potential frustration if the final purchase price approaches half a million dollars, though he admits he would likely still proceed. His wife, who does not share his enthusiasm for rapid acceleration, will probably not join him for drives.

Google has released Gemini 4 Carbon, its latest frontier AI model, positioning it as a major step forward in coding, agentic task completion, and cybersecurity defense. The announcement highlights Carbon's top ranking on the Artificial Analysis Intelligence Index (v5.0), where it outperforms leading rivals including Claude Opus 5.5 from Anthropic and GPT-6 Astra from OpenAI. The model builds on Gemini 4 Argon, which Google released earlier the same month, and represents the company's most capable model to date.

Gemini 4 Carbon is initially being rolled out exclusively to a select group of cybersecurity defense companies, with plans to expand access through Google's Fairwind Program over time. Google describes Carbon as reflecting the rapid pace at which it is advancing AI capabilities. The company says the model delivers a significant leap in cyberdefense performance, making it substantially more effective at identifying software vulnerabilities and producing secure code compared to its predecessor.

Carbon currently holds the top position on two specialized security benchmarks. The first, SEC-bench Pro, measures an AI model's ability to detect real-world software bugs. The second, ExploitGym, evaluates how effectively a model can transform an identified vulnerability into a functional exploit. These rankings signal that Google is targeting professional security researchers and enterprise-level cybersecurity teams as primary users, rather than positioning Carbon as a general-purpose consumer tool.

One of Carbon's notable design goals is its capability to sustain long, complex tasks without abandoning them prematurely. Google reports that its internal teams have seen the model run continuously for days, persisting through dead ends to complete demanding multi-step engineering challenges. The model is also engineered to avoid taking shortcuts, resist prompt injection attacks—a common method of manipulating AI systems—and collaborate more effectively with users throughout extended workflows.

Carbon also aims to be more transparent and honest in its outputs. Google claims it reduces unsupported claims by a factor of two compared to Argon, helping users rely more confidently on the information it provides. Despite its advanced capabilities, Carbon launches at the same introductory pricing as Argon: $2 per million input tokens and $10 per million output tokens. This competitive pricing suggests Google is aggressively pursuing market share in the rapidly evolving frontier AI landscape.