GenAI, a tsunami for the tech world
By Xavier Lazarus, with help from Justine Guers & Louisa Mesnard
Disruption by generative AI: what impact on startup development models?
The outburst of generative AI into our daily lives, thanks to, in part, the widespread availability of ChatGPT, can’t help but question for most of us the supremacy of human intelligence over the machine. From Descartes to Lévi-Strauss, language has always been considered to be mankind’s “quintessential cultural fact”.
But there are other languages, and in particular, one that is much easier for machines to understand and reproduce, because it is more standardised and less subject to subjectivity: computer language. Generative AI, which is both a code and a code generator, will therefore impact both development methods and the technological products themselves.
What impact can generative AI be expected to have on startups’ development processes?
After over 20 years of investing in European tech, I believe we’re on the cusp of a revolution that will shake up the developer profession and, in turn, the way technology companies operate. As with the arrival of the mobile internet, it’s difficult to predict the scale of the coming tsunami. But one thing is certain: entrepreneurs must adapt to this new paradigm, by seizing the opportunities generated by this wave (better productivity and integration within the offering), but also by acting quickly to avoid the pitfalls along the way (market changes).
Let’s take two typical examples of these changes within the Elaia portfolio:
- Mirakl, the SaaS e-commerce solutions provider, has integrated GenAI innovations in partnership with OpenAI to automate product referencing, resulting in major productivity gains for its customers, who can now use their technical resources on more strategic subjects without jeopardising the quality of the product data offered on their platform. The use of AI within Mirakl products, reinforced by this partnership, enables operators to: improve their conversion rate by 15%, integrate sellers on their platform 10x faster and reduce their customer support requirements by 90%.
- Arsen develops a platform for simulating phishing attacks. Over and above the obvious benefits in terms of content marketing and improved productivity for its developers, generative AI has radically changed the market in which Arsen operates.
Arsen has to deal with a new sophistication of attacks, which thanks to LLM are moving to conversational phishing models at scale, more difficult to detect both by the machine and by humans. This represents a major paradigm shift, which could lead to the widespread use of deep fakes and voice cloning, opening the door to increasingly sophisticated attacks (such as ultra-realistic telephone attacks). In this context, Arsen’s services must adapt and integrate generative AI within the solution to reproduce these attacks. Finally, agility is still on the human side, much more than on the machine: human vigilance adapts faster than the development cycles of protection solutions against these new threats.
How can tech entrepreneurs adapt?
First of all, you need to review your roadmap in the light of the new possibilities created by this revolution, but also of the likely threats linked to the emergence of GenAI. Secondly, you need to rethink the organization of your technical teams, using the new possibilities of generative AI in the field of development (i.e. automating and optimizing code generation), and repositioning the skills of technical teams where they have the greatest impact (in particular the creation, but also subjects linked to the management of private data, for example, whose use by LLMs can be viewed as sensitive).
Is this the end of the developer’s profession?
I doubt it, but the developer job is bound to evolve. Added value will continue to come from the design of solutions, and will be less linked to the ability to deliver code, which is likely to become a commodity.
In the same way, the advent of calculators didn’t put an end to mathematics, and the widespread use of computers didn’t destroy all service sector jobs; on the contrary, with sufficient hindsight, the arrival of these major innovations has pushed back the boundaries and enabled us to go further.
Computer code will become more universal, easier to generate thanks to GenAI, because it will be within the reach of non-experts, without the need for extensive training, and no longer at the expense of the user.
You can also find the 🇫🇷 version of this paper in Maddyness here.
In a 2nd chapter, I’ll be looking at the challenges that this revolution sets.