A year after Val Kilmer’s death, his digital ghost haunts the screen in As Deep as the Grave. The indie film’s trailer dropped at CinemaCon on April 15, showing an AI-generated Kilmer as Father Fintan, a priest blending Catholic rites with Native American spirituality. Archival footage, photos, and voice recordings trained the models. Once assets load, shots render in minutes. Kilmer’s family approved. Yet questions linger. This feels like a threshold moment. A person’s likeness outlives them as an asset.
Filmmakers defend the choice. Director Christopher Voorhees realized his vision without Kilmer shooting a frame, thanks to generative AI. The actor appears in a significant part—77 minutes of the runtime. Variety called it resurrection. Today.com noted the AI version’s seamless cadence. But critics brand it terrifying. Consent from the deceased? Family sign-off suffices now. Hollywood watches closely. Doug Liman’s Killing Satoshi slashed costs from $300 million to $80 million using AI tools, per X posts from TheWrap. Steven Soderbergh defends AI filmmaking in Variety. Reese Witherspoon embraces it too, via Hollywood Reporter.
Shift to code. Anthropic’s Claude Mythos Preview stunned even its creators. This frontier model spots zero-day vulnerabilities like no other. It found flaws in every major browser, OS, and the Linux kernel. A 27-year-old bug in critical security infrastructure? Uncovered. Thousands more during tests. Outperforms humans in cybersecurity tasks. Anthropic’s system card, released April 7, details it: substantial leaps in software engineering, reasoning, computer use. Anthropic deemed it too risky for public release. Instead, Project Glasswing shares it with 40+ partners—Apple, Amazon, Microsoft, Google—to patch holes before hackers strike.
Why the lockdown? Mythos exploits vulnerabilities autonomously. Cybercriminals could weaponize it. “This is Y2K-level alarming,” one expert told The Guardian. Anthropic CEO Dario Amodei briefed the White House, per Mercury News. The NSA uses it anyway, defying Pentagon tensions, reports Engadget citing Axios sources. Meanwhile, Anthropic shipped Claude Opus 4.7—weaker on cyber skills than Mythos but upgraded for vision, coding, presentations. Benchmarks place it between Opus 4.6 and Mythos.
Google’s Gemini fights back. Recent tests show its writing evades AI detectors better than ChatGPT or Claude. Tools flag rivals; Gemini slips through. TechRadar highlighted this. Gemini 3.1 Flash TTS delivers expressive speech. Notebooks organize chats, files into knowledge bases syncing with NotebookLM. A Mac app launched too. DeepMind’s ER-1.6 aids robotics reasoning, per their blog.
And the fallout spreads. Morrisons cuts hundreds of office jobs in an AI push, says The Grocer. LinkedIn execs deny AI causes hiring drops, but patterns emerge. Bank of England simulates AI amplifying market crashes, via Reuters. Ex-programmers warn of data center strains—drained reservoirs, power grabs. Stanford’s 2026 AI Index tracks the frenzy, at HAI.
Claude for Work betas in Teams, Enterprise. Managed Agents wipe startup pitches overnight, per Instagram buzz. OpenMythos reconstructs Mythos architecture openly—770M params matching 1.3B transformers via recurrent depth, on GitHub. Uber burns annual AI budgets on Claude Code already, The Information reports. Compute crunches hit: WSJ cover on energy shortages, here. Anthropic inks Google, Broadcom deals. Claude users gripe it’s dumber—AMD confirms nerfs.
Two worlds collide. Kilmer’s AI eyes pierce from trailers. Mythos probes code unseen. Gemini writes undetected prose. Jobs vanish. Markets tremble. AI accelerates. Partners hoard the sharpest tools. Public gets echoes. Threshold crossed. What’s next?
WebProNews is an iEntry Publication