journalfront

Inflection AI CEO: No Longer Competing on Next-Gen AI Models

2024-11-26

Just last year, Inflection AI was at the peak of startup excitement, presenting best-in-class AI models it claimed could surpass those of OpenAI, Meta, and Google. But now, as revealed by its new CEO to TechCrunch, the startup has shifted its focus. Microsoft hired the previous CEO, Mustafa Suleyman, to manage its AI business and acquired most of Inflection's staff and technology for $650 million. Since then, Inflection has been making significant changes. It started limiting usage on its consumer AI chatbot, Pi, and pivoted more towards enterprise customers. On Tuesday, it announced the acquisition of three AI startups in the past two months to enhance its offerings for global enterprises. The company also remains open to licensing AI models from its former competitors in the future. The Federal Trade Commission is investigating Microsoft's partial acqui-hire to assess potential competition issues. According to Inflection's new CEO, Sean White, who took over after the deal, the startup is no longer competing in building the next-generation AI models but still holds its ground in the enterprise front. "I am not going to, and don't feel the need to, compete with a company that is trying to build the next 100,000-GPU system," said White in an interview with TechCrunch, referring to the well-funded companies like Microsoft that can build frontier AI models. He clarified that while they can't compete in making the next-generation model, they are still competing with them in the enterprise space. "Our solution and tools are designed to meet the enterprise needs," he added. White believes today's AI models are sufficient for most enterprise needs and is skeptical about test-time compute scaling being the next generation of AI models. He thinks AI labs have cleverly rebranded high latency as "thinking" to make consumers feel better. Instead of focusing on the cutting edge of AI research, Inflection is now aiming to offer practical AI tools for enterprises. It announced the acquisition of Jelled.AI, which uses AI to manage employee inboxes, and BoostKPI, which offers AI data analytics tools. Last month, it acquired Boundaryless, an automation consulting firm in Europe to expand its overseas presence. White says Inflection is still using its own models but is open to using other models in the future. One of Inflection's value propositions is that its AI can run on-premise, unlike those from leading AI labs that need to be run in the cloud, which is appealing to enterprises concerned about data security. These acquisitions have strengthened Inflection's talent and product portfolio. However, it will face tough competition on the enterprise AI front. Salesforce has been investing heavily in AI agents, and Meta recently launched a new business AI unit. Startups like Anthropic and Cohere are also continuously building products for business customers. But Inflection believes it is better positioned to compete in the enterprise space rather than competing with frontier AI labs to make more capable models.

Inflection AI's Shift from Model Competition to Enterprise Focus

Inflection's Past Glory

Just a year ago, Inflection AI was the talk of the town. Its AI models were touted as the best in class, capable of outshining those from industry giants like OpenAI, Meta, and Google. This was a time of great promise and excitement for the startup.However, as time passed, the landscape changed. Microsoft's acquisition of Inflection's CEO and staff marked a significant turning point. The startup had to reevaluate its strategy and focus.

The Pivot to Enterprise

After the acquisition, Inflection began to shift its attention towards enterprise customers. It started limiting usage on its consumer chatbot, Pi, and redirected its efforts towards providing more tailored solutions for businesses.This pivot was not an easy decision. The startup had to let go of some of its initial ambitions and focus on what it did best - serving the enterprise market.

Acquisitions for Growth

In the past two months, Inflection has made a series of acquisitions to strengthen its position in the enterprise AI space. By acquiring Jelled.AI, BoostKPI, and Boundaryless, the company has expanded its product portfolio and talent pool.These acquisitions have allowed Inflection to offer a more comprehensive suite of AI tools and services to its enterprise customers. It has also helped the company stay competitive in a rapidly evolving market.

Competing in the Enterprise

Despite the shift in focus, Inflection is still competing in the enterprise space. Its new CEO, Sean White, believes that their solutions and tools are well-suited to meet the needs of enterprises.While they may not be competing in building the next-generation AI models, they are focusing on providing practical and effective AI solutions that can drive business value.

Value Proposition and Competition

One of Inflection's key value propositions is its ability to run AI on-premise. This gives enterprises more control over their data and security.However, the startup will face intense competition from other players in the enterprise AI market. Salesforce, Meta, Anthropic, and Cohere are all vying for a share of the business.But Inflection believes that its focus on the enterprise and its unique value proposition give it a competitive edge. It is well-positioned to succeed in this challenging market.

Amazon Web Services (AWS) is gearing up for its most significant conference of the year, with all indications suggesting it will be a momentous occasion. AWS re:Invent 2024 is set to commence in Las Vegas on Sunday and will continue through December 6. A keynote address by AWS CEO Matt Garman is scheduled to take place on December 3, starting precisely at 8 a.m. PT.

Key Livestream and Coverage

AWS has plans to make a live stream of the keynote available on YouTube. We will be closely following and reporting on all the announcements made on stage throughout the course of re:Invent. Stay tuned for these updates as well as our regular stories that will be emerging during the event.

New Leadership at AWS

This year marks Garman's first re:Invent as AWS' CEO. His predecessor, Adam Selipsky, stepped down in May after a three-year tenure. Garman now faces the challenge of addressing the market perception that AWS lags behind Microsoft and Google in the generative AI race. In recent months, AWS has introduced several generative AI services, claiming that its revenue from this area has reached "multiple billions" of dollars.

AWS has also achieved a significant win by strengthening its partnership with Anthropic. It became Anthropic's primary model training partner and agreed to have Anthropic's team assist in guiding its future AI chip design efforts. These partnerships and initiatives are crucial in AWS' pursuit of staying competitive in the AI space.

However, AWS has not been without setbacks. The departure of top AI exec Matt Wood and delays in its flagship Olympus model have presented challenges. Garman will likely address these issues during his keynote, as the re:Invent catalog lists over 500 generative AI-related sessions.

Innovation Across the AWS Stack

In an interview in October, Garman emphasized that re:Invent would not be "just an AI conference" this year. He stated that AWS will have innovations across the entire stack, including compute, storage, databases, analytics, and AI. This broad approach showcases AWS' commitment to providing comprehensive solutions to its customers.

Expect a special appearance by Amazon CEO Andy Jassy during Garman's keynote. The theme of the keynote, as described by Amazon, is "highlighting how AWS is reinventing foundational building blocks and developing new experiences to empower customers and partners in building a better future."

On December 4, AWS VP Swami Sivasubramanian will deliver a presentation focused on new AWS services related to AI and data. And on December 5, Amazon CTO Werner Vogels will delve deep into his tech predictions for the coming year, a regular and important part of re:Invent.

All of these speeches will be streamed on YouTube, and we will be providing comprehensive coverage right here on TechCrunch. Stay tuned for more updates and insights from AWS re:Invent 2024.

On Monday, Apple's selection of finalists for its highly regarded "iPhone App of the Year" award once again showcases how the iPhone manufacturer is minimizing the influence of AI technology on the mobile app ecosystem. Just as it did in the previous year, Apple's 2024 list of top iPhone finalists leans towards more conventional iOS apps. These apps assist iPhone users in performing specific tasks such as recording professional video with Kino, customizing their running plans using Runna, or organizing their travels with Tripsy. Notably, AI apps like ChatGPT, Anthropic's Claude, Microsoft Copilot, and those dedicated to creating AI photos or videos were not nominated for this prestigious award.

Surprising Exclusion of ChatGPT

Given the widespread popularity of ChatGPT, which has now become an Apple partner for Siri improvements, it is quite astonishing to observe that the app has not received any official year-end accolades from Apple's App Store editorial team. Despite introducing clever new features in 2024, such as an Advanced Voice Mode for interacting with the AI virtual assistant and a web search feature that competes with Google, ChatGPT has been overlooked. While ChatGPT is frequently featured in the App Store and Google Play's editorial suggestions, both Apple and Google refrained from nominating it as the overall winner last year. In fact, ChatGPT achieved remarkable success in 2023 when it became the fastest-growing consumer application, reaching 100 million users shortly after its launch. This year, Google designated the party-planning app Partiful as its app of the year.

AI-Powered Apps on Other Finalist Lists

Although ChatGPT was snubbed, a small number of AI-powered apps made appearances on Apple's other 2024 finalist lists, including those for the iPad and Mac App of the Year. Among Apple's list of 45 finalists across various categories in apps and games, AI-powered apps were only mentioned a few times. Moises, an app that provides AI tools for practicing music, was nominated along with the kids app Bluey: Let's Play and the animation app Procreate Dreams for the iPad App of the Year. Adobe Lightroom, which now incorporates AI-powered features, was nominated alongside the productivity app OmniFocus 4 and the 3D design app Shapr 3D for the Mac App of the Year.

AI in Cultural Impact Finalists

Among a dozen Cultural Impact finalists, only one app was described by Apple as being enhanced by AI technology - the language-learning app EF Hello. While other apps may utilize AI secretly, like the finalist Pinterest, their App Store marketing does not promote them as "AI" apps to consumers. This indicates that Apple's curated list of finalists emphasizes apps that empower human creativity rather than those focused on AI automation. Most of the finalists' apps enable users to do more with their iPhone or other devices without relying on an AI assistant or features. Whether it's designing, organizing, filming, creating, or playing, these apps offer unique value.

Apps for Human Creativity

For instance, Kino, the professional video app developed by the makers of the pro camera app Halide, aims to enhance the video-making skills of every user. It provides tools and features that help users become better videographers. On the other hand, there are apps that assist Apple's devices in functioning as productivity tools for those in creative fields. This year, Apple also introduced the new Apple Vision Pro category, where it will select both an app and a game winner. These finalists highlight the importance of human creativity and the diverse ways in which apps can enhance the user experience on Apple devices.