The Latest AI Breakthrough: What the Bottleneck Debate Means

MIT Technology Review recently covered a startup called Subquadratic that claims to have solved a major speed and cost problem in large language models. The company says it has found a way to cut the number of calculations these AI systems need, which could make them faster and cheaper to run.

If true, this would be a big deal for any business using AI. Right now, running powerful models costs a lot of money and uses a lot of energy. A breakthrough that slashes those costs could open the door for smaller companies to use advanced AI without breaking the bank. But experts are still skeptical, and the technology needs more proof.

Why Efficiency Matters More Than Hype

The real story here isn't just about one startup's claim. It's about a growing push across the AI industry to make models more practical for everyday use. Many businesses have held back from adopting AI because of high costs, slow response times, and energy consumption.

If companies like Subquadratic can deliver on their promises, we could see a shift from “bigger is better” to “smarter is better.” That would benefit Australian small and mid-sized businesses that don't have huge IT budgets. But scepticism is healthy. Many past AI breakthroughs have failed to live up to the hype. Business leaders should watch this space carefully and wait for independent verification before making big investments.

What This Means for Australian SMBs

For Australian SMBs, cheaper and faster AI means more opportunities to automate tasks like customer support, data analysis, and content generation. Instead of needing expensive cloud credits, a local retailer or accounting firm could run AI tools on standard hardware.

However, the technology is still emerging. Australian businesses should not rush into buying new AI systems based on claims alone. The key is to stay informed and test new solutions on a small scale first. That way, you can see real benefits without risking your budget.

What You Can Do Now

  • Audit your current AI spend to understand which tasks cost the most in compute and energy.
  • Follow independent benchmarks and third-party tests before adopting any new “breakthrough” AI model.
  • Start with a small pilot project using a lower-cost or open-source model to see if it meets your needs.
  • Train your team to evaluate AI tools critically, focusing on speed, cost, and accuracy metrics.
  • Work with a trusted IT partner to design a flexible AI strategy that can adapt as the market evolves.

Navigating AI changes can be complex, but you don't have to do it alone. MS&VG helps Australian SMBs cut through the noise and choose the right AI tools for their business—without the hype.