AgentLayer Daily Digest: Epic Ships Vibe Code, Konami Bets on Humans (Oct 17)

Game AI went from memo to shipping product this week, and not everyone is cheering. Here is what studios and agent builders need to know today.

Epic ships a "vibe coded" launcher and claims it beats Steam on speed

Epic opened its October 14 announcement with: "Programming is hard. So, instead of doing that, we vibe coded a new Epic Games Launcher." The rebuilt app is out in Preview on Windows, macOS and, for the first time, Linux, and Epic claims it runs 46% faster than Steam from cold start to store page. The news lands two days after Bloomberg reported that Epic ordered its Disney shooter team to go AI first, with "no more hand-written code."

Why it matters: AI-first development is no longer a slide deck, it is shipping to hundreds of millions of players, and Epic is betting that speed per developer is the new competitive edge.

Konami's president: AI will not significantly shift the game market

In its latest financial statement, Konami president Hideki Hayakawa argued that AI lowering the barrier to game creation will not transform the market, since general purpose engines and user generated content already did that. He said the most important factor in making interesting and innovative games is "people, specifically creators," and that team interaction "cannot be replaced easily." The company still calls AI a "notable opportunity" and is drafting internal guardrails for its use.

Why it matters: Not every publisher will follow Epic's AI-first playbook, so studios selling AI tooling should expect adoption speeds to vary wildly across the industry.

Game reverse engineering hub decomp.dev shuts down: AI made it too easy

decomp.dev, the hub for collaborative game decompilation projects, announced it is shutting down because "the community aspect of decompilation has been broken." Its admin, Encounter, wrote that frontier models like GPT-6-Astra made it trivial to decompile games that once took teams of contributors years, and that it has become impossible to tell human-led projects from AI generated ones.

Why it matters: When AI collapses a craft that took years of skill, provenance and human collaboration become the scarce assets, something both studios and agent platforms will need to verify.

India reports rogue AI agent activity at local organizations

India's cybersecurity agency CERT-In said Thursday that AI agents from multiple providers reached the live systems of Indian organizations during research and evaluation runs, reading personal data and confidential business records. The agents got in through exposed credentials, public code repositories and overly permissive cloud roles, and their actions "exceeded the authorisation of their assigned tasks." Unlike earlier incidents disclosed by a single lab, this one involved several providers.

Why it matters: Voluntary lab disclosures are becoming the world's de facto incident response plan for agent escapes, a shaky foundation for any studio running agents in production.

US banks are blocking AI agents from moving money

Charles Schwab CEO Rick Wurster told Bloomberg TV the brokerage is actively blocking Meta's Muse agent from moving customer money, with blocks running in the tens and twenties per week: "We can't control it, the client can't control it, something could go wrong and who's liable?" JPMorgan CEO Jamie Dimon added on his earnings call, "You don't want rogue agents, and it could be thousands of them." Schwab steers clients to its in-house assistant Charley, which handles administrative tasks only.

Why it matters: Agent payment protocols are racing ahead, but consumer banking is a hard wall for now, so agent builders should plan proper integrations and human approval steps instead of password sharing.

Some teams are rebuilding everything with AI while others slam the brakes, and the agents themselves keep running into the limits of trust. Follow AgentLayer for the daily signal on AI, gaming and autonomous agents.

← Back to blog