La start-up chinoise d'IA a dévoilé un modèle très performant avec un coût minime par rapport aux modèles américains. Une prouesse qui remettrait en cause les dépenses faramineuses des géants de la tech dans les puces.

Avec la troisième version de son grand modèle de langage, lancée en décembre, l’entreprise d’intelligence artificielle chinoise DeepSeek pourrait bien avoir battu en brèche l’idée que les meilleures IA requièrent des milliards de dollars et les puces les plus avancées. Pour seulement 5,6 millions, soit 7 % du coût du développement de ChatGPT, Pékin est parvenu à créer une IA comparable aux modèles américains.

In 1935, Austrian-born physicist Erwin Schrödinger described a thought experiment that magnified a glaring problem at the heart of quantum mechanics. To this day, the problem remains, summed up by Schrödinger's seemingly ludicrous notion of a cat existing in an undecided state of life and death.

The fastest way to destroy a great business is to turn it into a venture moonshot

En général, les arts et la culture sont particulièrement ouverts aux nouvelles approches, ce qui les rend plus enclins à expérimenter par rapport à d’autres secteurs. Pendant la pandémie de Covid, de nombreuses institutions culturelles ont profité de cette période pour explorer l’impression 3D, la production sur site et l’utilisation de matériaux écologiques. Cet esprit d’innovation ne se limite pas aux grandes scènes : l’impression 3D et les technologies associées ne sont pas seulement adoptées par de petits établissements en quête d’innovation, mais ont même fait leur entrée à Broadway !

L’impression 3D est souvent associée aux plastiques et aux métaux, mais aujourd’hui, nous nous intéressons à un domaine bien différent : la nourriture. Certaines imprimantes 3D permettent désormais de créer des sucreries, des pâtisseries, des chocolats et même des plats. Ces appareils offrent une grande liberté et créativité en cuisine, et peuvent, dans certains cas, aider à réduire le gaspillage alimentaire. Bien qu’elles ne remplacent pas encore les fours ou micro-ondes, elles représentent une avancée intéressante pour tous ceux qui aiment cuiciner. Aujourd’hui, nous nous intéressons aux imprimantes 3D alimentaires disponibles à la vente.

Regulatory agencies of any kind know that instituting new rules often requires companies to make substantial changes to their business. The ripple effects can fan out endlessly throughout every corner of the organization, with dependencies and contingencies that disrupt even the most experienced change management gurus. But regulatory agencies aren’t intending to incite upheaval, turmoil or failure with strict deadlines or the potentially complex adjustments required for compliance. Those deadlines to comply are put in place for a reason, and eventually, the grace period extended to companies between the time a law is approved and the time it’s implemented will come to an end. For many sustainability and environmental, social and governance (ESG) regulations, that time is now.

Gen AI has several attributes that humans lack. It’s always on. It boasts encyclopedic knowledge. It generates output instantly. It can scale endlessly. This new era of AI can feel intimidating for the limited, human life forms that created it. Humans now need to look harder to see where our unique value still lies. What tasks or skills might be difficult for AI for the foreseeable future? What moat can humans continue to own and defend? In this article, the author describes four strongholds for humans, and the associated human skills that remain valuable — for now.

While AI has the potential to solve major problems, organizations embarking on such journeys of often encounter obstacles. They include a dearth of high-quality data; too many possible solutions; the lack of a clear, measurable objective; and difficulty in identifying whether a proposed solutions is “good.” In tackling these issues, Google DeepMind, Alphabet’s AI lab that is striving to solve extremely challenging real-world problems, has come up with solutions that can benefit others. And it also offers guidance in choosing which opportunities to pursue after potential projects have overcome those hurdles.

The year 2024 was a challenging year for sustainability, from climate issues to inequality. Three top concerns include elections and political turmoil that threaten progress, companies retreating or staying silent on DEI and ESG goals, and the rise of sustainability reporting that’s (temporarily) taking up companies’ time. Other issues from the past year include the clean economy hitting tipping points, AI’s growth threatening decarbonization, heavy industry making some progress, and the beginning of a crackdown on “greenwashing.”

As businesses increasingly focus on innovation and customer-centric strategies, customer co-creation has emerged as a powerful way to involve customers in the product development process. By actively engaging customers in idea generation, design, and decision-making, companies can better understand market needs and create products or services that directly meet customer expectations. In this article, we’ll explore different customer co-creation models and how they help drive innovation.

We all like to think our best ideas are the product of our own brilliance. But if we’re being honest, some of the most innovative solutions don’t come from a single person—they come from a collective. The truth is, your best ideas might not even be yours, and that’s actually a good thing. Whether you’re working in a federal agency, a local government, or any other public institution, leveraging the ideas of others—especially through collaboration and crowdsourcing—can take innovation to a whole new level.

From idea competitions and startup incubators to virtual whiteboards and design thinking workshops, there is no shortage of mechanisms for brainstorming what’s possible in an organization. It’s exciting to have so many ideas to work with.But what happens next? Eventually, you need to identify which solutions best fit your business needs. Ideation on its own isn’t enough to drive innovation outcomes—and could even be a waste of resources if not governed properly. To drive outcomes, you need disciplined portfolio management.

Venture building is a tried and tested method for organizations to discover, validate, and incubate new business models. Originating in the US in the 1950s with the first Corporate Venture Capital (CVC) activities, organizations initially used this approach as a way of diversifying their portfolios and investing in areas outside of their core business.

To help corporate entrepreneurs leverage the critical actions that will ultimately become the foundation of a corporate venture’s scaling effort. Based on the insights from over 200 innovation tracks, this practical guide provides a deeper understanding of what you need to do early on, to ensure long-term venture growth.

How many times have you proposed a new idea and been told, “We can’t do that?” Probably quite a few. My favorite memory of being told, “We can’t do that,” happened many years ago while working with a client in the publishing industry:

It’s time to revisit the idea of avoiding the word “no,” but this time, I want to approach it from a different angle. First, let me tell you about my friend Dr. Nido Qubein, a devoted husband and father, motivational speaker, entrepreneur, president of High Point University, and mentor to many, who came to this country when he was 17 with little more than $50 to his name. Almost 50 years ago, he attended High Point University, never dreaming that one day he would be president of the institution, a position he accepted in 2005.

Wherever ambiguity is paramount, wherever judgment is required, or wherever moral values are at stake, one must not expect ChatGPT to be the final arbiter. That is simply not what it is designed to do. It can be an input, but it cannot be trusted to be the final output.

Do you feel like you’re drowning in a sea of bad news? You’re not alone. We’re standing in the eye of a storm of war, political division, and endless layoffs. In times like these, why bother with innovation when we’re using all our energy to survive and make sense of things?

It’s often said by teams that “diversity is our strength.” We take for granted the idea that diverse teams bring more lived experiences, ideas, and solutions to the table. When asked, “How does diversity affect teamwork?” most leaders assume that teams composed of individuals from different backgrounds, cultures, and perspectives are more likely to approach problems from various angles and come up with innovative solutions. And hence most leaders assume that diversity is a source of greater performance.And while that’s true—it’s not as clear cut as we assume.