Just recently, OpenAI gave us reason to wonder what the limits of Artificial intelligence could be when it announced the release of its newest model, OpenAI o1. This innovation is said to be equipped with unprecedented reasoning abilities in the fields of science and programming. There was also the news of the Generative Cinema with Sora, and then also that of the customization on the GPT Store.
And with this new wave of innovations, I think we all now have cause to wonder if AI has limits, or perhaps we are just being slow to discover all of its capabilities.
This new AI model, OpenAI o1, pauses before responding and can solve complex tasks, particularly in the fields of science, technology, engineering, and mathematics (STEM), using a “chain of thoughts” reasoning like a human would.OpenAI also says that the model has excelled in high-level tests, beating the GPT-4.0 in doctoral-level scientific subjects in biology, chemistry, and physics.
Tekedia Mini-MBA edition 15 (Sept 9 – Dec 7, 2024) has started registrations; register today for early bird discounts.
Tekedia AI in Business Masterclass opens registrations here.
Join Tekedia Capital Syndicate and invest in Africa’s finest startups here.
The tweet announcing this new model stated; “We’re releasing a preview of OpenAI o1—a new series of AI models designed to spend more time thinking before they respond. These models can reason through complex tasks and solve harder problems than previous models in science, coding, and math.”
The company described this AI as the ‘champion of code’ having reached the 89th percentile in Codeforces programming contests. It also excels in mathematics, correctly solving 83% of problems during an International Mathematical Olympiad test, compared to 13% for GPT-4.0.
Shocked? Just hold on for a little more information.
The new model has upped the game on cybersecurity, too, with a security test score of 84 out of 100 in resistance to “jailbreaking” attempts, compared to 22 for GPT-4.0. This is the result of a new training effort that makes it better able to adhere to security and alignment rules even in complex situations. And here’s one more. It is about 80% cheaper than the full version with OpenAI 0.1-mini, a cost-efficient reasoning model that excels at STEM, especially math and coding.
I think we can all agree that OpenAI did their homework on this one, but I am almost as sure that they have something bigger planned for the near future. But let’s begin to think about these innovations and the implications for the industry.
First off, I like the enhanced cybersecurity feature because, as I have always said, we need to ensure that we secure the data that people share with businesses online. As the news already stated, this feature strengthens OpenAI collaborations with AI institutions in the USA and the UK.
In terms of how these models will work well within the programming and coding space, we will have to see about that. I understand from the news that the OpenAI 0.1-mini coded a video game from a prompt, even though the company said it is optimized to improve the daily lives of developers by automating complex processes.
There is the question of whether it would reduce the need to hire actual human engineers. As this AI is said to have advanced coding capabilities, some pundits have already noted that its ability to pass OpenAI’s research engineer hiring interview for coding might be the final nail in the coffin for coding engineers.
While I understand the concern, I don’t see that it would work without getting the right input and prompts from someone who knows his onion in the trade. No doubt, the AI is capable of advanced reasoning, code generation, and debugging, but it still needs humans to play their part. What I think might happen – if this model is all they say it is – is that a company might just hire one or two coding engineers instead of having a team of twelve.
So, it is more likely to be like a productivity enhancer. Coding engineers would still be needed, after all, I do believe that even the AI models are being built by engineers. However, this is the season to focus on becoming exceptional because if you can only do as much coding as can be done by an AI model, you may soon be out of a job.