As Cohere and Writer Mine The 'Live AI' Arena, Pathway Joins The Pack With A $10 Million Seed Fundraising Round
Pathway has just raised a $10 million Seed round to build live AI systems that, claims the company, think and learn in real-time as humans do.
As large enterprises grapple with how to incorporate AI into their platforms and processes, they have encountered a problem: Generative AI needs to have memory and its training data must be constantly updated for it to have any practical use. This area is now called ‘Live AI’ and a number of startups are working in the space, including Cohere and Writer.
The Pathway round was led by TQ Ventures, with participation from Kadmos, Innovo, Market One Capital, Id4 and angel investors. Another investor in Pathway includes Lukasz Kaiser, the co-author of Transformers and a key researcher behind GPT o1 from OpenAI.
Pathway’s offering includes what it calls ‘infrastructure components’ that power live AI systems, feeding on structured and unstructured data, meaning that enterprise AI platforms can make decisions on up-to-date knowledge. Customers so far include NATO and La Poste, the French post office.
Zuzanna Stamirowska, Co-Founder and CEO of Pathway, told TechCrunch over a call: “The way deep learning and LLM assistants are working, is that you take the training data and then you train models. But the question is, how to deal with knowledge, how to deal with memory? Right now an LLM acts like a bit like a very smart intern on the first day of his job, being offered a book to read. But they can’t really memorize it. Plus, it’s not live, it’s stactic.”
To remedy this, she said Pathway “enables developers to build a pipeline where they can feed-in live data into the AI systems. Right now we do it during the prompting stage of when you build LLM applications or Gen AI applications.”
More on Pathways and their $10 M Seed round on TechCrunch
Mike Krieger, From Instagram To Anthropic
Mike Krieger is the Co-founder and former CTO of Instagram and recently joined Anthropic as Chief Product Officer. He joined Aditya to discuss how he's navigated the repeated -1 to 0 phases of his career and how he went from founding one of the biggest social apps of the mobile generation to joining one of the most exciting companies of the AI generation.
Who Needs A Keyboard — Hume Just Made It Possible To Control A Computer With Your Voice
Hume AI, the empathetic voice model company, has just unveiled a demo of an integration between Anthropic’s Computer Use technology, and Hume’s Empathetic Voice Interface (EVI) tech.
The video shared by Hume of the demo working shows a user talking to their computer screen to set up a hands-off chess game with the Hume persona. The computer sets up the board, invites the user to make the first move, and in the end shows complete mastery of the board, the computer and the conversation as the chess game progresses through three moves.
This all happens with no user input — no keyboard, mouse, or other physical connection apart from some sultry AI voice chat. Voice control of a chess game isn't new but this goes much further than that.
On the face of it, the technology behind the demonstration is fairly well established by now. One model, Claude, handles the computer interactions through its ability to ‘see’ the screen through multimodal training and activates functions as though it were pressing buttons on the keyboard. The Hume model translates voice to text commands and feeds them to Claude, while also translating the text output of the computer AI into dulcet tones for the user’s ears.
"By integrating Claude with EVI, we've created something truly special. Claude's frontier natural language capabilities and personality complement EVI's expression understanding and empathy, so EVI can “act out” Claude’s responses and generate fluid, context-aware conversations that feel remarkably human,” says Hume co-founder, Alan Cowen.
More about Hume AI’s empathetic voice model on Tom’s Guide
Unraveling AI Bias: Principles & Practices
Join Dinesh Shetty as he explores the fascinating world of AI and the critical topic of AI Bias and the principles of avoiding bias, emphasizing the importance of AI Governance in identifying and addressing bias.
Alibaba Releases An 'Open' Challenger To OpenAI's o1 Reasoning Model
A new so-called “reasoning” AI model, QwQ-32B-Preview, has arrived on the scene. It’s one of the few to rival OpenAI’s o1, and it’s the first available to download under a permissive license.
Developed by Alibaba’s Qwen team, QwQ-32B-Preview contains 32.5 billion parameters and can consider prompts up ~32,000 words in length; it performs better on certain benchmarks than o1-preview and o1-mini, the two reasoning models that OpenAI has released so far. (Parameters roughly correspond to a model’s problem-solving skills, and models with more parameters generally perform better than those with fewer parameters. OpenAI does not disclose the parameter count for its models.)
Per Alibaba’s testing, QwQ-32B-Preview beats OpenAI’s o1-preview model on the AIME and MATH tests. AIME uses other AI models to evaluate a model’s performance, while MATH is a collection of word problems.
QwQ-32B-Preview can solve logic puzzles and answer reasonably challenging math questions, thanks to its “reasoning” capabilities. But it isn’t perfect. Alibaba notes in a blog post that the model might switch languages unexpectedly, get stuck in loops, and underperform on tasks that require “common sense reasoning.”
More on Alibaba’s ‘Open’ challenger on Tech Crunch
Five Ways to Scale Your AI With Data And Analytics | Gartner’s Gareth Herschel
By improving business processes, enabling better decisions and supporting innovation, data and analytics leaders can deliver tangible business outcomes for their organizations and stakeholders.
Learning to harness the power of AI goes a long way toward supporting this endeavor.
In a sneak peak of what’s to come at Gartner Data & Analytics Summit in Orlando, FL on March 3-5, 2025, Gartner VP Analyst Gareth Herschel will share the top ways that data and analytics leaders can scale AI in service of their D&A initiatives.
Meta Plans To Build A $10 Billion Subsea Cable Spanning The World, Sources Say
Meta, the parent of Facebook, Instagram, and WhatsApp, is the second-biggest driver of internet usage globally. Its properties — and their billions of users — account for 10% of all fixed and 22% of all mobile traffic. Meta’s investments into artificial intelligence stand to boost that usage even further. So to make sure it will have reliable infrastructure to support that business, Meta is taking the pipes into its own hands.
TechCrunch has confirmed with sources close to the company that Meta plans to build a new, major, fibre-optic subsea cable extending around the world — a 40,000+ kilometer project that could total more than $10 billion of investment. Critically, Meta will be the sole owner and user of this subsea cable — a first for the company and thus representing a milestone for its infrastructure efforts.
Sunil Tagare, a subsea cable expert (and pioneer in the space, as founder of Flag Telecom), who was the first to report Meta’s plans back in October, told TechCrunch that the plan is to start with a budget of $2 billion but as the project builds out that figure is likely to go up to more than $10 billion as the project extends into years of work.
Sources close to Meta confirmed the project but said it is still in its early stages. Plans have been laid out, but physical assets have not, and they declined to discuss budget. The expectation is that Meta will talk more publicly about it in early 2025, when it will confirm plans for the cable, including intended route, capacity, and some of the reasoning behind building it.
More on Meta’s subsea cable efforts on TechCrunch
Google Subpoenas AI Rivals After DOJ Forces Chrome Spinoff
U.S. regulators want a federal judge to break up Google to prevent the company from continuing to squash competition through its dominant search engine after a court found it had maintained an abusive monopoly over the past decade. CNBC's Deirdre Bosa reports on news from Google.
The proposed breakup floated in a 23-page document filed late Wednesday by the U.S. Department of Justice calls for sweeping punishments that would include a sale of Google’s industry-leading Chrome web browser and impose restrictions to prevent Android from favoring its own search engine.
A sale of Chrome “will permanently stop Google’s control of this critical search access point and allow rival search engines the ability to access the browser that for many users is a gateway to the internet,” Justice Department lawyers argued in their filing.
Thats all for today, however new advancements, investments, and partnerships are happening as you read this. AI is moving fast, subscribe today to stay informed.