Kyeson Blake Utley, Founder of Ghost Key Development
Kyeson Blake Utley, Founder of Ghost Key Development

Software is being built faster than ever, and the pressure to launch quickly may be leaving security as an afterthought. The effects of that are already visible. The FBI's 2025 Internet Crime Report recorded nearly $21 billion in losses from cyber-enabled crime, a 26% increase on the year before. The report also found that more than $11 billion of it was linked to cryptocurrency. Speed may win a product its launch date, but it offers no protection once attackers start looking for a way in.

AI is accelerating development further, allowing almost anyone to turn an idea into working software. But working software may not mean a resilient and protected one. Researchers at Georgia Tech confirmed 74 publicly disclosed vulnerabilities linked to AI-generated code as of April 2026, with 35 found in March alone. The true number could be considerably higher, since much AI-generated code may never be traced.

The risks are not limited to today. In draft guidance published in late 2024, the US National Institute of Standards and Technology set out plans to begin phasing out widely used encryption methods such as RSA from 2030 and disallow them by 2035, as quantum computing threatens to break them. Systems handling financial or personal data may suffer the most. Software built today without the future in mind could end up being outdated or exposed within a decade.

The problem partly lies in how organizations consider security. Too often, software is built first and tested for weaknesses only at the end, if at all. Budgets for prevention can be hard to secure, while money for recovery tends to appear quickly once a breach occurs. By then, the most costly mistakes may already be built into the system.

Kyeson Blake Utley, founder of Ghost Key Development, approaches software from the opposite direction. A self-taught cybersecurity engineer, he began in penetration testing, the practice of ethically hacking systems to expose their weaknesses, before moving into development. Ghost Key Development builds custom, security-first software, with a focus on post-quantum cryptography and Web3 and crypto platforms. "Most developers ask whether something works," Utley says. "I ask how I would break it."

That mindset shapes how Utley structures every project. He explains that work is divided into small phases, and as each one is completed, he and a colleague swap and try to attack each other's code before testing their own again. Problems are fixed before the next phase begins, rather than at the end. "It can add time to development," he says. "But it costs far less than discovering a vulnerability after launch." He adds that the approach still allows him to deliver within competitive timelines.

In Utley's view, some risks are better removed than defended. On one crypto wallet project he worked on, the system was designed without the traditional seed phrase, the recovery code users often store insecurely, relying instead on biometric or passkey access. Rather than patching a vulnerability later, the risk was eliminated entirely. "Cryptography rarely fails," he notes. "The way people handle the keys does."

Utley believes AI-assisted coding has made strong technical foundations more important, not less. He adds that developers may be able to build quickly with AI, but they still need to understand hosting, infrastructure, and security to keep what they build safe. "If you can't build it without AI, you may not be ready to secure it," he says. Even small oversights, such as an exposed access key, can cost organizations heavily.

According to Utley, organizations might need to rethink how they budget for security. Investing early in prevention could cost much less than paying to recover from a breach. "Too often, the money for prevention cannot be found until after something goes wrong," he says. For people in security-sensitive roles, he suggests the same principle applies to everyday tools, choosing protection over convenience wherever possible. "If someone wants to target you, make it as hard as possible," he adds.

Utley does not expect developers and security engineers to be one role. Instead, he sees a future where the strongest teams build with an attacker's mindset from the very first line of code. Getting there may require organizations to value resilience as highly as they value speed. "Build it as if someone is already trying to break it," he says. As software becomes faster to produce and harder to protect, that mindset may matter more than speed itself.