Hugo Carreira’s Post

Is there no differentiation among big tech AI companies or is no differentiation possible? Or both? The big man tweets...or "x's." This would make more sense if Meta hadn't open sourced Llama. That was one of the more remarkable decisions in recent tech history, completely upending OpenAI and Microsoft's strategy. Google has now opened source "Gemma," a version of their AI model. To win spend from open source, Big Tech AI models will need to find a way to differentiate. It's unclear how yet. All the while, startups will be building on open source... Also, I wouldn't lump Elon Musk in as a source of hope at the moment... https://buff.ly/48zFDQD

Microsoft backed artificial intelligence startup OpenAI made a pitch for business in Japan on Monday as it opened its first Asia office in Tokyo. "This is just the first step in what I hope will be a long-term partnership with the people of Japan, government leaders, businesses and research institutions," OpenAI CEO Sam Altman said in a video message. The startup, which has caused excitement among consumers since the launch of its ChatGPT generative AI chatbot in late 2022, is looking to grow new sources of revenue globally. Microsoft said last week it would invest $2.9 billion over two years in cloud and AI infrastructure in Japan, part of a wave of investment globally by U.S. tech giants. Credit: Sam Nussey Christopher Cushing via Reuters #japan #asia #openai #artificialintelligence #generativai #samaltman #microsoft #investment #chatgpt

#TWIXR *NEW* Article released by: Charlie Fink Read the full story here: https://lnkd.in/gGH-d2NE Tech Industry Funding Updates: NVIDIA-backed Together AI Raises $106 Million, Adaptive AI LLC Raises $20 Million, and Tavus Developer Platform Raises $18 Million 1. Together AI Raises $106M: - Nvidia-backed Together AI raises $106 million in a new funding round, doubling its valuation since last fall. - The company offers a platform to easily develop with and deploy open-source models and productionize them in enterprise-grade applications. 2. Former Google Engineer Faces Charges: - Former Google software engineer Linwei Ding charged with stealing AI trade secrets from the company. - He had been uploading confidential files to his personal Google Cloud for over two years. 3. Adaptive AI Emerges with $20M: - Adaptive AI emerges from stealth with $20 million, aiming to enable LLMs to learn from how a company’s own employees or customers interact. - The seed investment is being led by Index Ventures with participation from other investors. 4. Tavus Developer Platform Raises $18M: - Tavus Developer Platform raises $18 million for its new developer platform, Replica API, powered by its new Phoenix model. - The Series A funding was led by Scale Ventures, with participation by other notable investors. 5. OpenAI's Approach to Media Deals: - OpenAI continues to make deals with media companies, recently with French and Spanish language news content train ChatGPT models.

On #China’s own #GenerativeAI revolution - Building on top of Meta’s #Llama (mostly “open source”) can help local companies escape the race for scarce (or banned), expensive computing resources. Replicating this would keep Chinese players competitive while shifting the burden to entities able to follow Meta’s long term bet (Is there any other?) - and I wonder about the impact that upcoming local regulations could have on such dependency. Compare this with EU-based #Mistral’s efforts to reduce computational costs at the model’s layer (introducing efficiencies vs. Llama, apparently: https://lnkd.in/dC5nQaZV - most of it flies over my head but the conclusions are straight-forward enough). This guarantees independence from a geo-political standpoint but I wonder whether deploying on top of it will be convenient enough to resist the temptation… data does have gravity, and so do training resources or end-user applications. Can we already anticipate the commodities and bottlenecks of 2030? At the very least Zuckerberg seems to have formed a very strong hypothesis. (Unrelated? I find it insulting that LinkedIn dares including a "Rewrite with AI" button on every post. Why not take such intellectual suicide one step further to a tempting "Replace your profile with a bot" - very discouraging) https://lnkd.in/d8mwaGfQ

New Post: Microsoft made a $16 million investment in Mistral AI - https://lnkd.in/gPMbqfPQ - Yesterday, Mistral AI, a Paris-based AI startup working on foundational models, announced a new large language model that could rival OpenAI’s GPT-4, a chat assistant and a distribution partnership with Microsoft yesterday. But Microsoft and Mistral AI buried the news — or at least an important part. As part of the partnership, Microsoft is investing €15 © 2024 TechCrunch. All rights reserved. For personal use only. - #news #business #world -------------------------------------------------- Download: Stupid Simple CMS - https://lnkd.in/g4y9XFgR -------------------------------------------------- or download at SourceForge - https://lnkd.in/gNqB7dnp

The challenge OpenAI is facing is very deep and well known contradiction between "platform" (research/infrastructure) against development of the "application" layer. The design of AI platform was dominating in the original vision of nonprofit, this is where all scientists placed, but it needs money, a lot of money for computations. Therefore Sam brought the capital and the idea to implement well-known use cases and force #AI community to build applications around #chatgpt, collect feedback to improve "the platform", increase sales, speed up the capitalization of the company and fund the infrastructure. Nonprofit still remains "platform" and fully dependent from the fuel coming from applications and board (Ilya, etc) probably thinks it's moving in the wrong direction and shifting focus. It's just another reflection of The Innovator's Dilemma where new companies can stay focused and improve incrementally the technology which is the same kind of platform, versus OpenAI killing small startups making business on the application level. It's not healthy ecosystem. If OpenAI wants to survive as a company, they have to Open Source their platform and basic models without loosing control over "fine-tuned" proprietary models giving them advance against all competitors. Basically, to make quick transition from closed to open innovation and get thousands of scientists involved and helping with the "platform" excelling. Just like Meta is doing with LLaMA, slow but faster and sustainable in the long run as AI community is highlighting the way. https://lnkd.in/eQxGsiGQ

Why are big tech giving away their #LLMs for free? If the product is free! what do you monitize? -"The big losers in the commoditization of LLMs may ultimately be the current hot and distruptive AI startups - companies like OpenAI , Anthropic , Character.AI , Cohere , and Mistral AI . When the 5 largest companies on the SP500 start giving away your main product for free, a reckoning may be on its way." https://lnkd.in/djGCn7tB #Artificialintelligence #Artificiellintelligens #Ai

Yesterday, Microsoft confirmed its AI partnership and investment with French start-up, Mistral AI.   Co - founded by three experienced individuals in April 2023; Arthur Mensch, previously worked at Google DeepMind, and both @Guillaume Lample and Timothée Lacroix worked at the Meta Platforms (i.e., Facebook, Inc.), Microsoft heralded the company as an “innovator and trailblazer” at the forefront of building efficient and cost-effective AI.   Despite no financial details having been disclosed, Microsoft has confirmed the partnership involves a small investment, suggesting a much smaller investment than Microsoft’s multi-billion investment into Open AI. Since its inception, Mistral AI have raised a total funding of circa $528m (£416m) over various funding rounds, securing a value of around £2bn.   Yesterday, Mistral AI released their own chatbot version, called Le Chat, which created a level of interest that saw the volume of traffic temporarily shutting down the application - one to watch in the future as the start-up progresses! Mistral AI Microsoft Guillaume Lample Timothee Lacroix Arthur Mensch #ai #investment #microsoft #lechat #mistralai #artificialintelligence #partnership #chatbot #funding https://lnkd.in/es4Ve2Ve  

"OpenPipe, a Seattle startup that wants to make it easier and cheaper for companies to train and deploy large language models, announced a $6.7 million seed round. The company, part of Y Combinator’s summer cohort last year, lets developers build LLM models tailored to their specific use case. The idea is that smaller models can result in better performance and lower cost compared to using larger general purpose models such as OpenAI’s ChatGPT. OpenPipe is betting that companies interested in using generative AI might not need a chatbot that can answer any question on any topic — but rather something that understands their product line and company policies deeply to provide customer service or other capabilities." https://lnkd.in/d9iBg-zK.

Perplexity AI 🤖 Is it the next big thing in search - 🔍 . Time will tell . But sounds promising … Nvidia founder Jensen Huang uses it daily too 🙀💡 If you haven’t please give it a try - link below 👇 https://www.perplexity.ai/ Features ✅AI-powered search: Perplexity AI uses natural language processing and machine learning algorithms to understand and respond to user queries. ✅Search and discovery: Perplexity AI allows users to search for information, ask questions, and explore topics in-depth. ✅Guided search experience: Perplexity AI's Copilot feature provides a guided search experience, allowing users to explore topics in-depth and learn new things. ✅Personalization: Perplexity AI allows users to personalize their experience through their AI profile, adjusting settings like language, output format, and tone. ✅Collaboration: Perplexity AI enables users to collaborate with others through Collections and Threads, making it easy to share and work on projects together. ✅ Pro Search: Perplexity AI offers a Pro Search feature, which engages with users, fine-tunes its answers based on their needs, and delivers more accurate results. ✅File upload: Perplexity AI allows users to upload files, such as images, to use as prompts for search queries. totally 💯 loved ❤️ this ! ✅Citations and sources: Perplexity AI provides citations and sources for its answers, ensuring transparency and credibility. ✅Available on multiple platforms: Perplexity AI is available as an app on both iOS and Android devices, as well as on the web. Well done Team Perplexity Ai #LLM #OpenAI #PerpelexityAI https://lnkd.in/gywy5BUJ