The release of ChatGPT-5 overshadows OpenAI's prominence
In a much-anticipated launch, OpenAI unveiled its latest creation, GPT-5, a more reliable and efficient language model designed to revolutionise various sectors. However, the rollout has been met with a mix of excitement and disappointment, as GPT-5 falls short of delivering Artificial General Intelligence (AGI) as some had hoped.
While GPT-5 demonstrates notable improvements in efficiency, safety, and multimodal capabilities, it continues to struggle with key qualitative drawbacks, particularly generalization and error types common to previous models. Critics argue that these issues, despite being reduced, have not been fundamentally cured. The model still exhibits hallucinations, making confidently wrong factual assertions and committing various types of mistakes.
Paul Armstrong, the founder of TBD Group and author of Disruptive Technologies, notes that GPT-5's failure to overcome these core issues, despite significant gains in other areas, has led some experts to say it has "hit a wall" in advancing beyond these challenges.
The disappointment surrounding GPT-5 can be attributed, in part, to over-promising during the GPT-4 era, which may have led to increased expectations for ChatGPT-5. However, it's always prudent to price capabilities, not buy shares in AI theatre.
In the aftermath of the launch, many investors, founders, and boards have expressed their disappointment. Sam Altman, the CEO of OpenAI, publicly acknowledged a "bumpy rollout" and community frustration.
Looking forward, the next 90 days should focus on governance for AI. This includes considering future AI contract clauses, sandbox testing of GPT-5, and prevention of rework. Leaders should also keep models away from unobservable judgment calls until human oversight is designed to catch low-frequency, high-impact mistakes.
Despite the disappointment, there are meaningful productivity gains to be had with thoughtful integration of GPT-5. In fact, GPT-5's drop in long-form factual errors can lead to productivity gains in compliance reviews, policy drafting, financial analysis, and legal discovery.
Moreover, competitive positioning should assume an accelerating arms race in AI, with OpenAI not being the only player. Boards need a clear read on vendor incentives, a plan to separate marketing from measurable progress, and a mechanism to capture real gains without buying every promise.
In conclusion, while GPT-5 is a capable, more reliable model, it is not yet the fully baked AGI that some had hoped for. Strategic horizon setting benefits from sobriety, as GPT-5 has not achieved general intelligence, but meaningful productivity gains are possible with thoughtful integration. For those who prefer GPT-4's tone and creativity, there's even talk of a return. The AI landscape continues to evolve, and it's essential to navigate it with a clear understanding of what these tools can and cannot deliver.
Business leaders should be cautious about the potential of GPT-5 to deliver general intelligence, as it has shown improvements in efficiency but continues to struggle with key qualificative drawbacks. Financial gains can be realized through thoughtful integration of GPT-5, particularly in compliance reviews, policy drafting, financial analysis, and legal discovery, but it's crucial to separate marketing promises from measurable progress and avoid over-investment in AI.
