November 30th is ChatGPT’s 2nd birthday, and since its launch it has been through so many ups and downs, a norm for any startup. But it has also managed to become a leading tech company alongside long time big players. As of August this year, ChatGPT had reported over 200 million weekly active users, and is used in different industries.
In September, the ChatGPT website had a total 3.1 billion visits, with a huge user base around the world the more AI becomes more popularly used. OpenAI has a major upgrade, codenamed “Orion,” that will launch in December. Rumours say it will introduce an agent capable of managing complex tasks like coding and web browsing.
What Were The Biggest Updates In 2024?
In January, OpenAI launched ChatGPT Plus, a paid subscription offering faster response times and early access to new features. February introduced the “Memory” feature, which allowed ChatGPT to recall details from previous conversations, creating a more personalised experience for users.
In May, GPT-4o was launched, bringing faster responses and better understanding across multiple languages. July saw the introduction of SearchGPT, which provided direct answers to user queries instead of relying on links. Then in September, Advanced Voice Mode was introduced, making interactions feel more natural with improved voice options and conversational flow.
What Do Industry Leaders Think Of ChatGPT’s Journey So Far?
CEOs, founders and investors have stepped in to share their reflections, findings and learnings from the AI startup’s journey since its launch in 2022. Running an AI startup isn’t easy, but it can be successful, as we’ve seen with OpenAI. Here’s what experts think:
Our Experts:
- Dominik Angerer, CEO and Co-founder, Storyblok
- James Evans, CEO and Founder, Command AI
- Alex Rice, CTO and Co-founder, HackerOne
- Kit Cox, CTO and Founder, Enate
- Leonid Feinberg, Co-founder and CEO, Verax AI
- Graham Glass, CEO and Founder, CYPHER Learning
- Nairy McMahon, Founder and CEO, Radial Change
- Dr. Imad Riachi, CEO and Founder, Honu
- Kira Makagon, Chief Innovation Officer, RingCentral
- Peter Wood, CTO, Spectrum Search
- Adam Levine, Business Coach, InnerXLab
- Zuzanna Stamirowska, CEO and Co-founder, Pathway
Dominik Angerer, CEO and co-founder, Storyblok
“Even two years on, it’s worth remembering that generative AI is at a relatively early stage of its development. Chat GPT is an impressive piece of technology but, as more and more people have interacted with it, limitations have become apparent. It is not wholly accurate and can be tricked.
“Data scientists will warn you that AI is only as good as the data it uses to learn from. If your data is incomplete, biassed or simply wrong, AI will amplify these problems with potentially disastrous results. There is a reason that Meta’s own version of Chat GPT uses a much smaller foundational language model. Fewer parameters limits potential but also means less could go wrong. Therein lies the challenge for businesses and also a glimpse at the future for marketers.
“The best way to start using generative AI is on an incremental basis and always with human oversight. For example, in the automation of tasks – like Virtual Identity’s image sourcing – or as a starting point for creativity – for example, generating email copy which is then adapted. This will enable your business to put in place the policies and procedures for leveraging generative AI in a risk free manner.”
James Evans, CEO and Founder, Command AI
“In the mainstream, the success of ChatGPT has made GPT synonymous with AI. This dominance seems to translate into the B2B world: In a study published by venture capital firm a16z, enterprises using AI predominantly use OpenAI’s models to power their AI features and apps. OpenAI’s models are in production at more than five times the rate as second-place finisher Google.
“Is this dominance here to stay? The data doesn’t seem to suggest so. In a16z’s data, most companies are now at least testing multiple models to avoid vendor lock-in. The data suggests that enterprises are excited about finding OpenAI alternatives to help them save costs, get more control and avoid vendor lock-in.
“OpenAI used to be the undisputed leader, but is no longer alone at the top. Many models approximate GPT-4o’s level of quality now, especially for less complex queries.
“If you rely on one API to run your entire product (or a crucial feature), that vendor can harm your business. Even if they don’t raise prices or shut down, simple downtime affects your customer/user experience. While you may have a preferred model like GPT-4, it’s wise to have other APIs (whether those host the same models or not) to mitigate vendor risk.”
Alex Rice, CTO and Co-Founder, HackerOne
“Across cybersecurity, criminals often make the first move. And artificial intelligence is no exception. As AI becomes increasingly commoditised and sophisticated, malicious actors have begun developing custom AI bots designed to impersonate trusted individuals with alarming accuracy. These AI deepfakes can mimic writing styles, speech patterns, and even emotional cues, essentially creating a skeleton key that bypasses traditional security measures built on interpersonal trust.
“The implications of this trend are far-reaching and deeply concerning. Imagine receiving a message that appears to be from a colleague, supervisor, or family member, complete with their typical mannerisms and inside knowledge. These AI bots can engage in prolonged conversations, adapting their responses based on context and gathered information.
“This level of sophistication makes it incredibly challenging for even the most vigilant individuals to distinguish between genuine communication and AI-driven social engineering attempts. As a result, organisations and individuals must adopt a new mindset of constant verification and implement multi-factor authentication systems that go beyond traditional trust-based security protocols.”
Kit Cox, CTO and Founder, Enate
“As a tech company, it was important for us to take a pragmatic approach to ChatGPT from day one. It’s become a valuable part of how we work here at Enate. Our engineers use it to draft their first cut of code, and our creative teams lean on it for proofreading, editing, and sparking new ideas.
“We don’t view ChatGPT as a threat, but as a ridiculously efficient personal assistant. For us, it’s about giving our people the best tools to do their jobs brilliantly. As the tech evolves, we’ll keep finding new ways to put it to good use.”
Leonid Feinberg, Co-founder and CEO, Verax AI
“ChatGPT has been a great deceiver. When it was first released, the industry took its almost magical capabilities to mean that the age of AI is upon us — everyone, even non-engineers, would now be able to build a dedicated AI product, tailor-made to address their needs, the industry assumed.
“Unfortunately, since then we’ve learned that a lot of work has been done on chatGPT even outside the ai models and this work is very hard to replicate elsewhere. It turns out that it is even harder to replicate specifically for enterprise-grade solutions, which have enterprise-grade requirements.
“We also learned that you do still need software engineering skills to build AI-based software and it’s not an easier task vs traditional software. ChatGPT jump-started the age of AI, but two years later we have more realistic expectations about how quickly this age is going to progress and how long we still have until ai is a real commodity, available to all.”
Graham Glass, CEO and Founder, CYPHER Learning
“It’s been two years since ChatGPT burst into the public consciousness, putting AI in the hands of the masses. By making ChatGPT publicly available, OpenAI was able to gather momentum at a pace – gaining more downloads than any application that had come before. We have quickly moved on from people tinkering around asking ChatGPT to write them a poem to adopting GenAI as a copilot for our lives – and with intelligent agents the next big trend, AI’s role in our lives will become even more tightly embedded.
The future? Even more ease of use. Only yesterday we were told that if would-be AI adepts want to wring anything useful from the technology, they’ll have to learn complex prompt design. This conviction permeated to the masses, with 57% of workers saying they’d like to use AI more, but they need training to create effective prompts. Yet we are now seeing such barriers being stripped away. Increasingly, generative AI delivers value as a well-controlled, intuitive, naturalistic element of a user-friendly application – in fact, if your favorite major application doesn’t already have AI embedded, you better believe it will soon. The notion that using AI takes specialised skills will quickly fade.
Good thing, too, because the idea was a natural limiter on AI’s potential – and could have exacerbated already-troublesome digital divides. The computer era’s most powerful, highest-impact revolutions have been those almost anyone can access. As AI evolves to understand natural language, user intent, and human behaviour, a broader swath of humanity can tap its capabilities. The digital world grows more inclusive. Far from making digital divides worse, AI as it progresses has the potential to shrink them.”
Nairy McMahon, Founder & CEO, Radial Change
“For a business to leverage ChatGPT effectively, it’s important to maintain critical thinking. Trusting AI or LLMs to provide valuable insights is key, but so is knowing when to rely on human intuition and experience. This balance ensures that AI serves as a powerful tool rather than a sole decision-maker.
“The best outcomes are given by a partnership between humans and AI, where each plays to its strengths. Business leaders should foster a culture where AI is seen as an ally, not a replacement. This involves training teams to work alongside AI, interpret its outputs, and use it to enhance their own capabilities.
“At Radial Change this is the beginning of our enquiry. For now, we know that aligning with AI requires understanding of its unique way of “thinking” and using that knowledge to complement human skills. By recognising the differences and leveraging them appropriately, we can harness the full potential of AI while maintaining the critical human elements that drive creativity, empathy, and strategic vision.”
Dr Imad Riachi, CEO and Founder, Honu
“The biggest lesson learned on ChatGPT so far
“GenAI and the advancements introduced by some of the new tech giants, like OpenAI, have seized the world’s attention, sending shockwaves through businesses worldwide. It has brought mass awareness of AI and dramatic increases in productivity in specific areas, and propelled ‘AI Agents’ (also known as ‘Autonomous Agents’) to the center stage.
“Driven by a state of FOMO, many are hastily jumping on the AI bandwagon and scrambling to take action quickly before proper assessment of the claims put forward by the proponents of this new technology. The noise levels around artificial intelligence are overwhelming, and the promises around its applications are reaching saturation point and full of false claims.
“Now, as we’ve seen recently, we are heading to plateauing AI Agent capabilities. The biggest lesson we must head is that the market should not throw more money at more computing power or larger models and, instead, ask if alternative architectures are another way of reaching AI nirvana.
“AI’s next evolution
“There is a lot of magical thinking surrounding AI, with a pervasive belief among businesses that AI is a silver bullet, capable of solving all their challenges. Its growing sophistication has the potential to reshape organizational leadership, how decisions are made, and the roles of business stakeholders.
“Today’s agents excel in language processing and information retrieval tasks. However it is not yet capable of the rational decision-making needed to get to reach this place.
“While attempts have been made to build AI capable of strategic thinking and acting upon it, they still fall short of the goal not only due to LLM hallucinations, but also because of intrinsic limitations of the LLM paradigm in grasping business context and capability of reasoning.
“This is why we see autonomous agents as simply automation tools rather than real autonomy, taking on the workhorse role whilst human managers and directors remain the owners of the strategic decisions that determine what a firm does next.
“We started Honu with the mission of delivering full business autonomy and the promise of a Self-Thinking Business. We saw early on that deeper networks and more powerful chips were not the answer to unleashing AI’s full potential in businesses. Fast forward to now and that is what is occurring. So what next? We see the answer emerging from a reimagining of the intelligence stack to bridge this ‘cognitive’ gap and the creation of a digital enterprise nervous system.”
Kira Makagon, Chief Innovation Officer, RingCentral
“This month marks two years since ChatGPT launched, fundamentally shaping the way we interact with artificial intelligence (AI). In a short span of time, generative AI evolved from a promising innovation to a transformative force across industries, enhancing productivity, fostering creativity, and redefining the way we communicate and collaborate.
At RingCentral, communication and collaboration are at the heart of what we do, and AI greatly benefits how we do it. As we look ahead, the continued advancement of generative AI holds immense potential, particularly for leveraging voice data – specifically, the conversations people have with customers, partners, and prospects.
“Such data is the richest fuel for growth and, when properly harnessed, it drives innovation, enhances customer satisfaction, and accelerates business outcomes. AI helps us all, customers and employees, to enhance productivity and keep us focused on the tasks that AI does not help us do – yet.
“The future of AI promises even greater advancements in enhancing customer experience and customer value. In 2025, I believe AI will remain a powerful partner in driving success, especially in turning untapped data into actionable insights that foster innovation.
“A human-centred approach to such innovation will continue to be crucial as we all learn to balance enthusiasm for innovation with intentionality and specificity as we all transform to an increasingly AI-led paradigm in high-tech in 2025 and beyond.”
Peter Wood, CTO, Spectrum Search
“ChatGPT has redefined what it means to innovate, showing startups that success often lies in simplicity and user-centricity. By removing the barriers that typically make complex technology intimidating, it made AI more approachable for everyday users. This focus on accessibility has shown that cutting-edge solutions don’t need to be inaccessible or complicated to stand out. For startups, this is a critical takeaway: innovation thrives when it is intuitive and responsive to the needs of real people.
“In its early stages, ChatGPT also highlighted a crucial lesson about transparency. As its capabilities grew, so did questions about its inner workings. A lack of clarity around how it operated led to concerns about misinformation and bias. For any business, this stresses the importance of open communication. It’s not enough to simply deliver a product; you must also make it clear how it works, what its limitations are, and where improvements are needed.
“Moreover, the ability to respond quickly to changing demands and expectations has been vital in ChatGPT’s evolution. For startups, this underlines the importance of agility—not just in creating products but in constantly adapting to new information, user feedback, and market shifts. Mistakes are inevitable, but they should be used as stepping stones to improvement.
“As for applying ChatGPT’s potential in business, the takeaway is clear: AI is a tool that can enhance, not replace, human creativity. While it can efficiently handle repetitive tasks and help brainstorm ideas, it’s the unique insights and emotional intelligence of people that drive real innovation. Startups should look at AI as a collaborator rather than a replacement.”
Adam Levine, Business Coach, InnerXLab.
“ChatGPT’s rise shows that innovation is only part of the equation; focusing on user needs is just as crucial. By stepping into a space where quick, human-like exchanges were required, it captured both attention and market interest. For startups, this is a reminder that technology needs to adapt to real-world uses – not just follow a preset script. Flexibility is key, much like how Lego’s modular design thrives by evolving.
“However, ChatGPT’s journey also had its bumps. There were missteps that affected trust, but these offer a valuable lesson for startups: underpromise and overdeliver. Transparency is essential – users can forgive flaws, but they won’t tolerate dishonesty. Mistakes should be seen as learning opportunities, not setbacks. Addressing problems head-on and resolving them quickly builds stronger relationships.
“When it comes to using ChatGPT in business, it can enhance productivity but not replace human insight. For example, it helps organise data efficiently, allowing time for the deeper, more intuitive aspects of work. The speed of AI complements human creativity, making it a partnership that adds value rather than simply automating tasks.
“ChatGPT also highlights how AI fits into our ever-changing world. Its development reflects the unpredictable, evolving nature of technology. While AI challenges traditional approaches to problem-solving, it also serves as a tool that can improve our understanding – without replacing the foresight and creativity that drive progress. Good tools amplify what we do; they don’t do it for us.”
Zuzanna Stamirowska, CEO and Co-founder, Pathway
What has ChatGPT as a startup taught other startups?
Something that is clear from ChatGPT’s undeniable success over the past two years is that go-to-market is everything and commercialisation is key. We would never have seen such profound changes in the development and speed of adoption of artificial intelligence (AI) had it not been for the very simple, usable product that is ChatGPT. The simplicity of a chat box to input questions and receive answers sparked the imaginations of the masses – it was a golden strike.
The next step will be for OpenAI to repeat this success with their GPT o1 model, which is a breakthrough in terms of advanced reasoning abilities. However, it hasn’t yet been translated into a format that broadly captures hearts and minds in the same way as its predecessor.
