MythWorx CEO Jason Williamson was formerly VP of Oracle for Startups. [Photo: Oracle]

Backed by benchmark results and a newly appointed CEO, MythWorx is betting its human-inspired AI architecture can challenge the LLM status quo—and bring AGI closer to market.

The Dallas-based startup announced it has named Jason Williamson as CEO. Founded in 2023, MythWorx recently emerged from stealth mode with technology that Co-Founder and Chairman Wade Myers says was built on “27 years of development.” The company positions itself as an artificial general intelligence (AGI) innovator, describing its technology as a step beyond traditional large language models.

Wade Myers

Myers said Williamson’s appointment validates the company’s technology approach.

“Jason’s decision to join MythWorx is a powerful validation of the team and our technology that addresses the power usage and compute limitations of LLMs,” said Myers, a serial software entrepreneur and general partner at Eagle Venture Fund, in a statement. “His leadership experience at Oracle and AWS, combined with his visionary mindset, makes him the ideal leader to maximize our potential as the next generation of AI innovation.”

Williamson most recently served as executive director at Oracle, where he led global innovation strategy and enterprise technology initiatives. There, he helped grow the company’s startup customer base from 250 to more than 3,000.

Earlier in his career, Williamson held leadership roles at Amazon Web Services (AWS), where he launched and led the firm’s private equity and management advisory partnerships. A veteran of the U.S. Marine Corps, he brings experience across corporate, startup, and government sectors.

“My focus is on growing the team, safeguarding and promoting our groundbreaking AGI innovations and benchmarking results, and delivering our capabilities to market,” Williamson said.

A compact model—and an AGI vision

MythWorx said its technology emphasizes real-time reasoning rather than sheer scale, in an approach the company frames as a move toward AGI.

In recent benchmarking using MMLU-Pro, an open-source evaluation tool that tests AI performance across dozens of academic and professional subjects, the company’s Echo Ego v2 model delivered what MythWorx called “standout results.”

According to the company, the model achieved 71.24% accuracy without any pretraining, retries, or chain-of-thought prompting. It answered questions in about 1.2 seconds per query, compared to the 5–10 seconds typical of mainstream LLMs.

Despite having just 14 billion parameters, the company reported that Echo Ego v2 outperformed much larger models, including DeepSeek-R1. The model achieved 87.64% accuracy on the math section and demonstrated strong domain-specific reasoning in fields like law, physics, and medical diagnostics—without fine-tuning, the company noted.

MythWorx emphasized the model’s energy efficiency as a key differentiator. The startup claims Echo Ego v2 uses roughly one-tenth the power of a typical large language model for the same workload—something the company sees as essential for long-term scalability and sustainability in AI.

Williamson said the results reflect the strengths of the company’s hybrid design.

“Other firms run on power-hungry GPUs in a highly inefficient way that simply cannot scale at the current pace,” he said. “They require massive datasets and pretraining to perform acceptably. We have a hybrid architecture that’s designed to self-improve and adapt just like a human brain, without the need for any pretraining.”

Williamson sees it as a turning point for AI development. “It’s a ‘black swan’ moment for AI,” he said, “that will democratize access to high-performance reasoning through elegant design, not raw power.”

Human-inspired AI architecture

According to MythWorx, its system uses what it calls “biomimetic learning,” an architecture designed to mirror human cognitive processing. Unlike traditional models that require vast datasets and tuning to improve performance, the company says Echo Ego v2 adapts and reasons in real time.

The model includes built-in moral constraints, according to the company, an attempt to align responses with human values around fairness, responsibility, and transparency.

MythWorx said it is now entering joint ventures and master licensing agreements to bring its system to market.

Lance Murray contributed to this report.

Don’t miss what’s next. Subscribe to Dallas Innovates.

Track Dallas-Fort Worth’s business and innovation landscape with our curated news in your inbox Tuesday-Thursday.

 

R E A D   N E X T

  • Data scientist Anmolika Singh put Dallas on the global AI Tinkerers map. At the first meetup, more than 30 pros—founders to Fortune 500 technologists—showed up to trade ideas, projects, and solutions.

  • Avoid common pitfalls with practical insights from guiding Fortune 500 companies through AI adoption

  • From Corporate Politics to AI, Lee Brown’s Research Transforms Education and Earns Him a Spot Among DFW’s Top Innovators.

  • AI seem overwhelming? Just go to office hours. 

  • Most people are getting AI all wrong—asking random questions and getting mediocre results, say Babar Bhatti and Masood Khan. Their new guide shows how to prompt like a pro and turn AI into your strategic team.