MythWorx appoints Jason Williamson CEO as Echo Ego v2 outperforms larger LLMs

MythWorx names Jason Williamson CEO as it pushes a human-inspired AI architecture beyond today's LLMs

MythWorx has appointed Jason Williamson as CEO and is positioning its "human-inspired" AI architecture as a path to artificial general intelligence. The Dallas-based startup says recent benchmark results show its compact Echo Ego v2 model can rival much larger systems while using less compute and energy.

Co-founder and chairman Wade Myers called Williamson's move a signal of confidence in the company's 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," he said.

Why executives should care

The company reports performance and cost advantages that matter to P&L owners: faster responses, lower energy draw, and reduced dependency on massive pretraining runs. If validated, this could change budgeting for AI projects, edge deployment, and compliance workflows that require predictable reasoning and auditability.

The headline numbers (company-reported)

• Echo Ego v2: 14B parameters
• MMLU-Pro accuracy: 71.24% with no pretraining, no retries, no chain-of-thought prompting
• Math section: 87.64% accuracy
• Latency: ~1.2 seconds per query vs. 5-10 seconds typical of mainstream LLMs
• Energy use: ~10% of a typical large language model for the same workload
• Outperformed larger models, including DeepSeek-R1, according to MythWorx

The leadership signal

Williamson most recently led global innovation strategy at Oracle, helping grow its startup customer base from 250 to 3,000+. Prior to that, he launched and led private equity and management advisory partnerships at AWS. A U.S. Marine Corps veteran, he brings cross-sector experience across enterprise, startup, and government.

"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.

Architecture: real-time reasoning over brute force

MythWorx frames its approach as "biomimetic learning," aimed at mirroring human cognitive processing. The company says Echo Ego v2 adapts and reasons in real time and includes built-in moral constraints to align outputs with values like fairness, responsibility, and transparency.

Williamson argues the model's hybrid design reduces the need for massive datasets and pretraining. "Other firms run on power-hungry GPUs in a highly inefficient way that simply cannot scale at the current pace," he said. "We have a hybrid architecture that's designed to self-improve and adapt just like a human brain, without the need for any pretraining."

Go-to-market: partnerships and licensing

MythWorx says it is pursuing joint ventures and master licensing agreements to commercialize the system. For enterprise buyers, the value prop centers on total cost of ownership, latency-sensitive use cases, and domains requiring domain-specific reasoning (e.g., law, physics, medical diagnostics).

What this could mean for your roadmap

• TCO planning: If performance-per-watt holds, you may be able to redeploy budget from GPU capacity to integration and risk controls.
• Edge and on-prem: Smaller models with low latency open options for regulated environments and data-local architectures.
• Safety and governance: Built-in moral constraints are promising; insist on transparent policy controls and override mechanics.
• Benchmark rigor: Request third-party validation of MMLU-Pro and domain tests, including prompts, sampling, and inference settings.
• Vendor durability: Assess IP posture, funding, and support guarantees before committing critical workflows.

Due diligence questions to ask MythWorx (or any compact reasoning model vendor)

• Can independent labs reproduce the reported MMLU-Pro results under standardized settings?
• What are the exact hardware, energy, and latency profiles under real workload conditions?
• How do the model's "moral constraints" work, and how are they audited for bias and failure modes?
• What is the update cadence, rollback plan, and telemetry provided for enterprise monitoring?
• What licensing terms cover IP, indemnity, and data retention across cloud, on-prem, and edge?

What to watch next

• Third-party evaluations and red-team results across high-stakes domains
• Pilot case studies with measurable ROI and incident reporting
• Evidence that the hybrid, human-inspired design maintains performance as features and users scale

MythWorx views the moment as a break from scale-at-all-costs thinking. As Williamson put it, it could democratize access to high-performance reasoning through elegant design rather than raw compute.

If you're building executive AI fluency and capability across teams, explore role-based programs and certifications at Complete AI Training.