Latest news with #InflectionAI
India.com
02-07-2025
- Business
- India.com
Meta owner Mark Zuckerberg spends Rs 1170000000000 to hire this man, his name is.., his expertise is to...
Mark Zuckerberg is making major moves to establish Meta's role in the fast-changing world of artificial intelligence (AI). In a big deal, he has partnered with a major player in the AI world who could help push Meta's AI aspirations further. To finalize this arrangement, it is reported that Zuckerberg invested a whopping $14 billion. According to the media reports that Meta has acquired a company named Scale AI. However, the truth is that Meta simply made a large investment in Scale AI-it didn't acquire the company. If it was an acquisition, then Meta would have had to buy all shares of Scale, and all employees would have had to either receive Meta stock or cash out in some manner. This did not happen. Instead, Meta invested $14 billion in Scale AI, which raised the valuation of Scale to $29 billion, and then made Meta's stake nearly 49%. Scale is still a separate company, and its board was unchanged. However, where there is this level of influence, it is very likely that the company is going to fall directly in line with Mark Zuckerberg's vision. Alexandr Wang is the founder and CEO of Scale AI, and he was critical to making this deal happen. Wang is joining Meta but will remain on the board of Scale. With Meta and Wang's combined stake, they have potential control of Scale AI now. To put it another way, for the foreseeable future, key decisions for Scale AI could practically be dictated by Meta. The deal was so large that some thought of Meta as buying the company outright. In fact, a significant amount of the capital ultimately went to Scale AI's employees because they were able to cash out their shares-partially, of course-but retain, and cash in, a percentage of their ownership. This allowed them to profit immediately while also staying invested in the company's future growth. It's said that this idea came from Alexandr Wang himself, ensuring that his team benefitted alongside him and didn't get left behind. The most interesting thing about this acquisition is that it seems like Meta is not truly interested in Scale AI's core businesses. Scale AI is primarily a data labeler—providing the prep work for training machine learning models, which is usually a human-intensive task. It is also a low-tech task and therefore low in innovation. Scale works a lot with big clients such as Toyota, General Motors, and various governments, who want to adopt AI, except have no idea how to build AI. For Meta, a tech of its size, Scale's business does not seem to quite fit either. Meta is not building a B2B data service business, and Scale's datasets are not valuable enough as datasets to warrant a deal on that level. The real purpose for the deal, it seems, Meta wanting to acquire Alexandr Wang, the CEO behind Scale AI. This is not unprecedented. Google invested in Character AI and lured some of their best employees onto their Gemini team. Microsoft did something similar with Inflection AI. So why is Alexandr Wang so significant? In the modern tech race, the player that builds the strongest large language models (LLMs) will win the game. It is a race to claim market territory. There remain many who claim they can build LLMs, but success is impossible without the right data, enormous compute, and the ability to scale. Users will always go with the highest-performing model. When it comes to this game, second best doesn't matter. Meta has not kept pace in the AI race to date. OpenAI has already claimed the consumer software market with ChatGPT, and Google and Anthropic are established developer players. Meta has models made like Llama 2, but they have not been able to put the flag in the ground claiming 'first' in what is becoming a heated market. To this point, Meta's play has been to keep it open-source, and this was enough to gather a broad audience of developers and researchers. Now, Meta understands open source can take them only so far. They need a visionary leader capable of defining their AI future; in this case, Alexander Wang is expected to be that leader. Meta is falling considerably behind in the AI race. OpenAI has taken the consumer space using ChatGPT, and Google and Anthropic have taken the developer space. While Meta has developed some models like Llama 2, its unable to stake a claim to the top of the competitive landscape. Meta's approach thus far has been to keep everything open-sourced, and that did help garner a large community of developers and researchers,. Nevertheless, the company now realizes it cannot simply rely on open source. They need a lossy visionary leader to mold their AI future, which is why Alex Wang is in the limelight.
News18
01-07-2025
- Business
- News18
Why Mark Zuckerberg Spent Rs 14 Billion To Get Alexander Wang To Meta
Last Updated: Meta relied on open-source to attract developers, but now seeks a visionary leader to shape its AI future—prompting the $14B bet on Alexander Wang to lead the charge Mark Zuckerberg is reportedly under pressure as Meta struggles to keep pace in the rapidly advancing world of artificial intelligence. In a bold move to change course, Meta has made a massive investment aimed at strengthening its AI capabilities. The tech giant has reportedly poured $14 billion into Scale AI, a leading data-labelling startup, effectively doubling the company's valuation to $29 billion. The deal is said to give Meta a significant 49% stake in Scale AI—along with a strategic edge in the AI race. Despite the substantial investment, Scale AI remains an independent entity with no changes to its board. Nevertheless, Meta now wields considerable influence over the company's operations. Alexander Wang, Scale AI's founder and CEO, plays a pivotal role in this arrangement. Although Wang retains his position on Scale's board, his partnership with Meta means the tech giant effectively steers Scale AI's decisions. The deal was substantial enough to create the impression that Meta had acquired Scale AI entirely. In reality, a significant portion of the deal benefited Scale AI's employees, who received substantial payouts for their shares while retaining some equity. This arrangement, reportedly Alexander Wang's idea, ensured that his team could profit from the company's growth. Why Is Meta Interested In Scale AI's Business? Meta's interest in Scale AI is particularly noteworthy, given that the latter's primary business involves data labelling for machine learning, a service with minimal technological innovation. Scale AI caters to clients such as Toyota, General Motors, Etsy, and various governments, providing data preparation services for those keen on adopting AI but lacking the in-house capability to develop it. This investment in Scale AI does not align with Meta's core business interests, as Meta is not looking to become a B2B data service company. The primary objective of the deal was to bring Alexander Wang into Meta's fold, a strategy similar to Google's investment in Character AI and Microsoft's acquisition of talent through Inflection AI. The Race To Build The Best LLM In today's AI-driven world, the company that builds the best Large Language Model (LLM) will dominate. It's a battle for market leadership, where knowing how to build models isn't enough. Without the right data, massive computing power, and the ability to scale, survival is unlikely. Meta is currently trailing in the AI race. OpenAI has dominated the consumer space with ChatGPT, while Google and Anthropic hold strong positions in the developer ecosystem. Although Meta has released models like Llama 2, it has yet to secure the top spot in the LLM race. Meta's core strategy so far has focused on open-sourcing its models, which helped attract developers and researchers to its ecosystem. However, the company now believes that open-source alone isn't enough. What it needs is a visionary leader to steer its AI future—and that's where Wang comes in. He is seen as the ideal choice to take Meta's AI ambitions to the next level. First Published: July 01, 2025, 18:55 IST
Mint
01-07-2025
- Mint
Can chatbots really understand you? Meet Pi, an emotionally intelligent AI that wants to listen and help
For years, AI chatbots have been racing to sound more human, but most still struggle to move beyond robotic politeness or scripted empathy. Now, another contender is stepping into the spotlight: Pi, an emotionally intelligent chatbot from Inflection AI, designed to make conversations feel less like talking to a machine and more like confiding in a thoughtful friend. It tries to offer a more personal and empathetic experience, positioning itself as a different kind of AI companion compared to popular tools like ChatGPT. Pi isn't just another chatbot for answering trivia or writing emails. Its core mission is to support users emotionally, offering gentle prompts, reflective questions, and a listening ear. The interface is inviting, blending artistic visuals with a choice of expressive voices, giving every conversation a touch of personality. Unlike many AI tools that focus on productivity or technical tasks, Pi is all about well-being. It can help users process a tough day, practice giving difficult feedback, or simply check in with their feelings - Pi claims to help with all. The interface includes options for voice interaction and AI-generated art, which add some personality and creativity to the experience, but the core interaction remains text-based. Pi's approach to conversation is refreshingly different. Instead of bombarding users with information or advice, it listens, remembers past chats, and responds with empathy. The chatbot encourages self-reflection, nudging users to think deeper about their emotions and goals, rather than rushing to solve problems. Voice interaction is a standout feature, adding warmth and nuance to each reply. Pi also generates creative art to accompany chats, making the experience feel more engaging and less transactional. While ChatGPT is known for its versatility, handling everything from coding to content creation, Pi carves out its own space as a companion for emotional support and personal growth. Here's how they compare: Aspect Pi (Inflection AI) ChatGPT (OpenAI) Main strength Emotional intelligence, support Technical tasks, creativity Conversation style Warm, reflective, supportive Informative, adaptive Features Voice, art, journaling prompts Coding, research, plugins Best For Self-reflection, tough talks Productivity, content, coding Pi doesn't try to outdo ChatGPT on technical grounds. Instead, it focuses on what many users crave: a safe space to talk, reflect, and be heard. For users seeking detailed, task-oriented AI help, Pi's limited feature set may be a drawback. The thrill of using Pi comes from its ability to make digital conversations feel genuinely caring. For anyone tired of chatbots that feel like digital assistants, Pi offers a new kind of interaction - one where the AI listens, remembers, and responds with genuine warmth. It's not for everyone. If you need a research assistant or want to build an app, Pi won't compete with the likes of ChatGPT or Gemini. But for journaling, practicing hard conversations, or simply having a gentle check-in, Pi is a breath of fresh air. Privacy is another big draw. Pi has scored well in independent studies for keeping user data safe, which is increasingly important as more people turn to AI for personal support.
Tom's Guide
30-06-2025
- Business
- Tom's Guide
I just tried the emotionally intelligent chatbot that's trying to pull you away from ChatGPT
AI has come a long way in its conversational skills. Those who have been using it since day one will remember the robotic replies that ChatGPT would muster up with a slight twinge of personality glued onto it. However, the technology has come a long way since then, with the vast majority of AI chatbots now able to throw their weight behind a friendlier, more conversational tone. And yet, AI is clearly still struggling to strike that friendly tone effectively. But one AI company believes they've mastered the art, offering an emotionally intelligent chatbot that really gets you. Pi isn't exactly new to the world of AI. It was launched just over two years ago by the company Inflection AI. It made our list of the best ChatGPT alternatives as the most personal option and recently scored highly in a study of the best AI chatbots for privacy. However, despite its determination to become the most emotionally intelligent chatbot, Pi never really took off, with one of its founding members leaving the company a year ago, taking almost all of the workforce with him. Get instant access to breaking news, the hottest reviews, great deals and helpful tips. While this sounds like the end for the company, it essentially merged with Microsoft, getting a huge round of funding and going under Microsoft's wing. It also had a new CEO come in, and one of the other co-founders is still involved with the project. Pi is still available today, and after its strong score for privacy, and people's concerns around other chatbots' strange attempts at emotional stability, what better time to pull it back out and see how it matches up? Pi looks like a cross between Claude and ChatGPT, offering artistic drawings, a variety of pre-designed prompt situations, and stylish text (that looks very similar to Claude). When you sign in, you have to pick a voice from a list of options. While all robotic, they do have some personality to them. The chatbot itself operates similarly to its competitors. Either ask a prompt with the chatbox, or use a pre-selected option. Where it differs is its more conversational tone and use of a voice answering all of your questions along with text. I spent time chatting to Pi about world news, fun facts about the ocean and even my day. While it is clear that the model is still an AI (obviously) it does succeed at hitting a balance between friendly but not annoying. The auto-generated art also looks great, and somehow feels more personal than most AI-generated art. The conversational tone is especially effective with more negative prompts. Where ChatGPT will overload you with questions, Pi acts more like a friendly voice to talk to. In all conversations there is a more natural flow. Pi isn't the same as other AI chatbots, and that's both a good and bad thing. If you want a detailed marketing plan, a scan of the latest trends, or to make an app or anything vaguely advanced, Pi falls behind pretty much every other AI app. In fact, even on simpler prompts, you'll get a far better response from the likes of ChatGPT, Claude, or Gemini. It also can't generate images, video, or code, features that most AI chatbots are now offering. Even more importantly, you can't upload files, and while the model can access the internet, it doesn't offer deep research like some of its competitors. It works best as an AI journaling service or a system to bounce your feelings and ideas back between Instead, Pi is a stripped-back service, even more now that it isn't getting the same push or funding that it did in its early days. It works best as an AI journaling service or a system to bounce your feelings and ideas back between. Because the conversational tone is more natural, it is useful for practicing hard conversations, like giving bad news or going for a job interview. It also works well for breaking down complicated ideas in a fun and engaging way. More importantly than all of that, and possibly Pi's biggest selling point, is that it is completely free. It is also likely to remain that way for the meantime. According to the company's blog, they are continuing to update the model, bringing in new features. It could one day offer a lot of what makes the likes of ChatGPT better, but without losing its leading tone.
The Hindu
20-06-2025
- Business
- The Hindu
OpenAI and Microsoft: A partnership under strain
A 'head over heels' relationship between a tech titan and an AI startup that began over six years ago is turning sour. Microsoft and OpenAI's pact powered the startup's artificial intelligence engine to build generative pre-trained models and de-aged the software maker for the AI era. Now — after cumulative investments swelling to $13 billion — the couple is battling between mutual reliance and burgeoning autonomy. This recalibration carries weighty implications for both firms. Recent reports suggest Microsoft is prepared to halt discussions over the future contours of its OpenAI alliance if disagreements on critical terms — like Microsoft's future equity stake — persist. The Windows software maker would then rely on its existing commercial contract, ensuring access to OpenAI's technology until marks a potential inflexion point in a relationship that saw Microsoft's capital and cloud infrastructure propel OpenAI to the vanguard of AI. Heart of the matter At the heart of the current negotiations are fundamental differences in strategic outlook. OpenAI has been overtly seeking to lessen its dependency on Microsoft for cloud computing, a move underscored by new partnerships. Notably, OpenAI finalised a deal in May to use Google Cloud's infrastructure, a significant step to diversify its computing resources beyond Microsoft's Azure —its current exclusive provider. It has also partnered with CoreWeave and is exploring arrangements with Oracle as part of Project Stargate to further expand its compute capacity. Such diversification provides OpenAI with technical alternatives and, presumably, greater negotiating leverage. The shifting personal ties between the firms' leaders, Satya Nadella of Microsoft and Sam Altman of OpenAI, mirror these corporate recalibrations. Once in near-constant communication, with Mr. Nadella reportedly texting Mr. Altman five or six times a day, their interactions have become more formalised, primarily consisting of scheduled weekly calls, per news reports. This devolution from spontaneous chats to structured exchange began after Mr. Altman's brief ousting from OpenAI in late 2023 — an event that led Mr. Nadella to rearchitect his company's AI future. While Mr. Nadella backed Mr. Altman, the Microsoft CEO also made his controversial decision to bring DeepMind's Mustafa Suleyman on board. At that point, Mr. Suleyman was running Inflection AI. And as part of the deal, the entire team at Inflection AI joined the software maker. Despite these undercurrents, public pronouncements remained diligently choreographed. Earlier this year, Mr. Altman posted a picture with Mr. Nadella on X, announcing the next phase of their partnership to be 'much better' than anyone is ready for. Mr. Nadella echoed the optimistic sentiment. Such displays were aimed at reassuring investors amidst intricate private negotiations and mounting competition from other AI players, as well as increasing regulatory scrutiny globally. A pivotal point A pivotal point of disagreement between the duo is OpenAI's corporate structure. In May, OpenAI announced it would restructure into a Public Benefit Corporation (PBC), while keeping its non-profit parent in control, retaining the authority to appoint board members . This was a significant shift from earlier considerations of a more conventional for-profit transition that might have diluted the non-profit's oversight and authority. The move, amidst criticism from OpenAI early investor and Tesla CEO Elon Musk, was aimed to better align its operational structure with its stated mission of developing AI for humanity's benefit, while still attracting substantial investment. This restructuring requires Microsoft's assent as a key stakeholder — with the tech giant having provided billions of dollars in funding. Microsoft is said to be negotiating the size of its own potential stake in this new PBC, with discussions reportedly ranging from 20% to 49%. Failure to finalise this restructuring by year-end could jeopardise funding from other investors, including a significant investment from SoftBank. Broader AI strategy Microsoft, for its part, is not standing still. Its AI strategy is visibly broadening beyond its OpenAI relationship. At its Build 2025 conference, Microsoft showcased integrations of models from Anthropic and Musk's xAI, signalling a move towards a more diversified AI portfolio. The company is also developing its own smaller, in-house models, like Phi-4, to reduce costs and reliance on any single provider for its Copilot services. This reflects a growing confidence in its proprietary capabilities and a desire to offer a wider range of AI tools on its Azure platform. Indeed, Microsoft's ability to leverage its existing agreement with OpenAI until 2030 offers it strategic latitude. But the evolving Microsoft-OpenAI dynamic unfolds against a fiercely competitive AI landscape. Both entities are balancing the fruits of their collaboration against the imperatives of strategic independence and market differentiation. Microsoft's potential willingness to pause talks and OpenAI's multi-cloud strategy both signal a relationship that is turning sour. The denouement of these negotiations will not only chart the future courses of the two firms but also establish significant precedents for partnerships, governance, and commercialisation in the rapidly maturing AI domain. The relationship, once a lodestar for AI collaboration, now offers a salient lesson in managing the intricate dance of shared ambition and diverging paths in an industry perpetually remaking itself.
