This Week in NLP #316
Keep up with what happened in NLP in the week ending Friday 6th December 2024.
Above the Fold
Overwhelmed? Here’s our pick for five things to know about this week.
Here’s TechRadar’s summary of everything announced at Amazon’s re:Invent 2024. [TechRadar]
Amazon and Anthropic are developing Project Rainier, a powerful AI supercomputer featuring Trainium chips. [Wired]
Apple surprised everyone with its presence at re:Invent. [Analytics India Magazine]
OpenAI plans to unveil new features, including the text-to-video AI tool Sora and a reasoning model, over 12 days starting December 5th. [The Verge]
Spotify Wrapped 2024 introduces an AI-powered personalized podcast recapping users’ listening habits, featuring interactive synthetic hosts and powered by Google’s NotebookLM. [Wired]
Now read on for everything else that happened in NLP this week.
Making News:
What makes this newsletter different?
This Week’s Topics:
If you’re reading this in a desktop web browser, you can access a navigation menu at the left-hand edge of this window.
Amazon re:Invent
At this week’s AWS re:Invent, Amazon introduced Amazon Nova, a suite of advanced foundation models, emphasizing high performance, cost-effectiveness, and industry-leading capabilities. [Business Wire]
Amazon is developing Olympus, an advanced multimodal AI system focusing on image and video analysis, to enhance its text processing capabilities and improve AWS cloud services. [The Decoder]
Amazon Q Developer enhances software development with AI-powered automation for unit testing, documentation, code reviews, and operational issue resolution. [Business Wire]
AWS announced new capabilities for Amazon Q Business, enhancing its generative AI-powered features to streamline and personalize enterprise workflows. [Business Wire]
Amazon SageMaker Unified Studio integrates analytics and AI tools within a single platform, enhancing data collaboration and reducing processing time for businesses. [Yahoo Finance]
Amazon SageMaker introduced new HyperPod capabilities and partner applications to streamline the development and scaling of generative AI and ML models, reducing costs and improving efficiency. [Yahoo Finance]
AWS introduced the Trn2 UltraServers and Trainium2 chips to enhance AI training and inference performance at lower costs, with future upgrades planned via Trainium3. [Yahoo Finance]
AWS announced updates to Amazon Bedrock, including Model Distillation for training smaller, customizable models and Automated Reasoning to detect and correct AI hallucinations. [VentureBeat]
AWS announced HyperPod Task Governance, a solution that optimizes AI costs by increasing GPU utilization up to 40%. [VentureBeat]
And Amazon is expanding Bedrock with a new AI model marketplace. [GeekWire]
AWS introduced new services to enhance enterprise AI by simplifying structured and unstructured data integration into retrieval augmented generation pipelines. [VentureBeat]
AWS and GitLab announced an integration of GitLab Duo with Amazon Q to enhance AI-powered DevSecOps, improving developer productivity, security, and innovation. [Business Wire]
AWS has updated Amazon Connect with new AI features for better segmentation, service automation, and performance measurement, enhancing contact center efficiency. [siliconANGLE]
Comet has partnered with AWS to integrate its AI model evaluation platform into Amazon SageMaker’s new partner AI apps. [Business Wire]
Informatica announced innovations for generative AI and analytics applications on AWS, including partnerships with SageMaker Lakehouse and support for AWS PrivateLink Resource Endpoints. [Business Wire]
New Relic and AWS have integrated their AI platforms to streamline workflows and automate incident response. [Business Wire]
Rubrik announced Rubrik Annapurna, an API service integrating with Amazon Bedrock, to enable secure data access and build generative AI applications customized to enterprise needs. [Business Wire]
ServiceNow and AWS have expanded their collaboration to integrate AI solutions, enhance automation, and optimize cloud investments, available via AWS Marketplace. [Business Wire]
Teradata is enhancing its collaboration with AWS to offer rapid-start Gen AI solutions using Teradata VantageCloud integrated with Amazon Bedrock. [Business Wire]
The Generative AI Wars
Apple and Baidu face challenges integrating AI features into iPhones in China, impacting sales amid Huawei's rising market share and privacy policy conflicts. [Yahoo Finance]
By June this year, China had 230 million generative AI users, with Baidu's Ernie Bot leading the market, followed by ChatGPT and Google’s Gemini. [South China Morning Post]
ChatGPT now boasts over 300 million weekly users. [The Verge]
Google has launched the Gemini AI app for Android and iOS, enabling Google Workspace users to leverage AI tools on mobile devices. [TechRadar]
Despite promoting its own Llama AI, Meta also utilizes OpenAI's GPT-4 in its Metamate tool to enhance functionality for developers. [Fortune]
Microsoft’s top AI customers include Adobe and Meta. [The Verge]
Microsoft is testing Copilot Vision in Edge, an opt-in AI feature for browsing that chats about website content. [Engadget]
And Microsoft acknowledged a bug in Windows 11’s Recall feature for Dev build testers. [TechRadar]
Nvidia is opening its first Vietnam R&D center to advance AI development, partnering with the government and local entities to boost the country’s tech ecosystem. [Yahoo Finance]
OpenAI is considering ads to cover AI development costs but emphasizes no active plans. [ITPro]
OpenAI introduced ChatGPT Pro, a $200 monthly subscription tier, offering enhanced features and access to advanced models o1 and o1 Pro Mode for improved reasoning and faster responses. [Ars Technica]
OpenAI hired three senior engineers from Google DeepMind to join its new Zurich office, focusing on developing multimodal AI models. [Wired]
Elon Musk’s attorneys have filed a preliminary injunction against OpenAI, alleging anticompetitive behavior. [TechCrunch]
And Musk’s xAI plans to expand its Memphis supercomputer to one million GPUs to compete with OpenAI. [Yahoo Finance]
The United States, China, and the United Kingdom are the most prepared to foster AI development, as per Stanford HAI’s comprehensive AI readiness ranking. [TechRepublic]
Feature Creeps
Meta is updating Threads with improved search and trending features amid increased competition from the rapidly growing platform Bluesky. [The Verge]
Google has introduced several new AI-powered features for Android and Pixel devices, including Expressive Captions for more dynamic closed captioning and enhancements to the Gemini chatbot and accessibility tools. [Wired]
Google's Gemini AI assistant now integrates with Spotify for voice-controlled music playback, currently available only in English. [The Decoder]
Google's NotebookLM AI podcast creator may soon be integrated into the Gemini app, enabling users to generate podcasts from PDFs and videos. [TechRadar]
Hype Bubble?
Jensen Huang, Nvidia's CEO, is promoting AI as a transformative societal force globally. [Wired]
OpenAI CEO Sam Altman suggests AGI will arrive sooner than expected but have less impact initially, with significant changes occurring later as superintelligence emerges. [The Verge]
Mira Murati, former OpenAI CTO, is optimistic about achieving AGI, while emphasizing responsible AI development and societal guidance. [Wired]
But TechDirt argues that generative AI’s lack of modularity makes it unreliable, untestable, not meaningfully open, and potentially a technological dead-end. [TechDirt]
Hardware
Dell is expanding its AI capabilities with new PowerEdge servers featuring AMD 5th Generation EPYC processors, enhancing performance, efficiency, and scalability for enterprise AI workloads. [TechRadar]
Italian startup iGenius and Nvidia plan to launch one of the largest deployments of Nvidia’s servers in Italy by mid-next year. [Yahoo Finance]
Intel’s CEO has been forced out by its board, frustrated with slow progress. [Bloomberg]
TSMC is in talks to produce Nvidia's Blackwell AI chips at its new Arizona plant, though final packaging will occur in Taiwan. [Yahoo Finance]
It’s Only a Model
01.ai's Yi-Lightning AI model claims sixth place globally with a US$3m budget, using 2,000 GPUs and efficient engineering despite restricted access to advanced hardware. [TechRadar]
Alibaba has released the Qwen with Questions (QwQ) model, an open-source LRM with improved reasoning capabilities over OpenAI's o1, licensed for commercial use. [VentureBeat]
Here’s Alibaba’s blog post on QwQ-32B-Preview. [Qwen]
DeepSeek's R1-Lite also challenges OpenAI's o1 in chain of thought processing. [ZDNet]
Google Cloud has launched new video and image-generation models, Veo and Imagen 3, on Vertex AI. [VentureBeat]
Google DeepMind's Genie 2, a successor to Genie, generates complex and interactive 3D worlds from images and text. [TechCrunch]
Luma AI announced the release of its updated Ray 2 Video Model for creating videos from text and images,. [Business Wire]
Microsoft has introduced industry-specific AI models and agents, developed through partnerships, to enhance customized solutions and drive innovation across various sectors like agriculture, automotive, and manufacturing. [eWeek]
Nexa AI's Omnivision is a compact vision-language model for edge devices, offering reduced computational requirements and strong performance in multimodal tasks. [InfoQ]
Nous Research is pre-training a 15-billion parameter LLM using a novel distributed method called DisTrO. [VentureBeat]
Nvidia's Hymba-1.5B-Base is a small language model with a hybrid architecture. [Analytics India Magazine]
OpenAI is expected to announce the release of its text-to-video model, Sora, during a 12-day event showcasing new features and products. [Engadget]
Prime Intellect’s INTELLECT-1, a 10B parameter model, demonstrates successful decentralized training across multiple countries. [Prime Intellect]
Runway introduced Frames, an AI image model offering greater control. [The Decoder]
Twelve Labs unveiled Marengo 2.7, a multimodal embedding model with a 15% performance improvement, utilizing a pioneering multi-vector approach for enhanced video analysis. [PRWeb]
Whose Data?
Amazon is reaching out to news publishers about opportunities to license their content for the next generation of the Alexa voice assistant. [Axios]
Canadian news organizations sued OpenAI for copyright infringement, accusing it of profiting from their content without permission, and seeking damages and usage bans. [Engadget]
A Columbia Journalism School study found ChatGPT’s citations are often inaccurate, posing reputational risks for publishers, regardless of their licensing agreements with OpenAI. [TechCrunch]
Dow Jones has negotiated AI usage agreements with nearly 4,000 news publishers. [NiemanLab]
Jeff Jarvis, a prominent US media critic, warns against exclusive AI partnerships between tech companies and major publishers, saying these deals shut out smaller publications and put journalism at risk. [The Decoder]
OpenAI is partnering with publisher Future to provide ChatGPT users access to content from over 200 media brands. [The Verge]
The UK is considering changes to its copyright laws regarding AI to balance protecting creators’ rights and allowing AI companies to innovate, potentially diverging from the EU’s approach. [TechRadar]
A survey reveals only 7% of UK authors have consented to AI training usage of their work, highlighting a need for fair compensation and licensing. [Publishers Weekly]
The LLM Ecosystem
Andrew Ng’s aisuite provides an OpenAI-like API for seamless integration and switching between various popular LLMs. [InfoQ]
Couchbase has launched Capella AI Services to enhance enterprise AI by improving data proximity, security, and integration. [VentureBeat]
Humane AI's CosmOS, an AI operating system for connected devices, aims to attract third-party adoption despite the company’s previous struggles with its AI pin and limited real demonstrations. [Engadget]
Liquid AI's STAR framework introduces a modular and automated approach to AI model architecture design, achieving significant efficiency and performance improvements over traditional Transformer models. [VentureBeat]
Pinecone has announced updates to its vector database, introducing a cascading retrieval approach and new reranking technologies to enhance accuracy and optimize enterprise AI applications. [VentureBeat]
Pixeltable launched its open-source AI data infrastructure to streamline AI development by reducing infrastructure complexity and compute costs. [Business Wire]
Pydantic launched PydanticAI, an open-source agent framework in beta, designed to simplify scalable, LLM-driven application development with features like type safety and modular workflows. [VentureBeat]
Qodo has released its autonomous AI regression testing agent, Qodo Cover, to simplify and improve software testing by generating and validating high-quality tests throughout development. [VentureBeat]
Reasoner claims to offer a cost-effective, neurosymbolic AI reasoning engine providing more accurate, explainable results for enterprise applications than existing LLMs. [Yahoo Finance]
SHIZA announced its Developer Suite, a no-code AI development platform optimizing performance and security for individual and enterprise use. [Business Wire]
Tata Communications is launching an AI Cloud to transform AI use in India, offering a comprehensive stack for enterprises, startups, and government with Nvidia partnership support. [Analytics India Magazine]
Traefik Labs launched the Traefik AI Gateway to simplify AI integration, enhance security, and improve observability for enterprises, while also integrating Treblle's AI API Assistant to boost developer productivity. [Business Wire]
Uber’s Prompt Engineering Toolkit centralizes prompt template creation and management, enabling effective and safe LLM utilization through features like evaluation, collaboration, and batch generation. [Uber]
Veritone's AI Solutions Group offerings are now on AWS Marketplace, allowing organizations seamless access and deployment of innovative, efficient, scalable AI services. [Business Wire]
Weaviate has launched Weaviate Embeddings, a SaaS service offering flexible and scalable open-source and proprietary data vectorization without rate limits or external dependencies. [GlobeNewswire]
Weights & Biases announced the general availability of W&B Weave, a toolkit supporting generative AI workflows, aiding developers with evaluation, monitoring, and iteration of AI applications. [Business Wire]
Wikimedia Deutschland is collaborating with DataStax to improve accessibility and usability of Wikidata for AI/ML developers, using advanced vector embedding technology. [Business Wire]
Agentic AI
AWS's Multi-Agent Orchestrator manages multiple AI agents, supports Python/TypeScript, and integrates with various environments. [InfoQ]
AI coding tool Cursor is adding autonomous coding agents in its latest update. [The Decoder]
Emergence AI unveiled its cross-application autonomous multi-agent AI orchestrator, emphasizing integration, flexibility, and enterprise-grade security to outperform existing solutions from Microsoft, Salesforce, and Amazon. [VentureBeat]
EPSoft.ai provides AI Agents that streamline tasks and enhance efficiency, offering businesses AI integration without technical expertise. [PRWeb]
Microsoft has released Magentic-One, a modular multi-agent system for complex tasks that enhances efficiency and addresses potential risks. [InfoQ]
Tech companies, including Perplexity, OpenAI, and Google, are developing AI shopping agents to automate online purchases, potentially transforming the online retail landscape. [TechCrunch]
Skyflow launched new features for Agentic AI, enhancing enterprise security and privacy in AI agent deployment through partnerships with Databricks and Workato. [Business Wire]
A new survey highlights AI agents powered by LLMs are transforming software interaction by automating GUI tasks through natural language commands. [VentureBeat]
And a report by Deloitte finds that by 2025, 25% of enterprises using generative AI are expected to deploy AI agents, with this figure projected to rise to 50% by 2027. [ZDNet]
Other LLM Sightings
Aktana introduced the Action Agent, an AI solution for life sciences, to enhance strategic omnichannel engagement and integrate with CRM systems like Veeva and Salesforce. [Business Wire]
Anjuna announced the release of Anjuna Northstar, an AI Data Fusion Clean Room enabling secure, collaborative AI-driven analytics for data integration. [GlobeNewswire]
Appriss Retail has introduced SecureGPT, a generative AI tool for its Secure platform, enhancing data analysis efficiency and usability for loss prevention specialists through natural language queries. [Business Wire]
Baidu has launched an AI assistant on its Ernie Bot platform to help generate and verify official documents in collaboration with China’s Xuexi Qiangguo app. [Yahoo Finance]
Cylerian's new AI Assistant offers real-time data analysis for Managed Service Providers and Managed Security Service Providers, providing critical cybersecurity insights and automating responses within their platform. [PRWeb]
Diligent has enhanced its One Platform with AI Smart Prep, Google Drive integration, and an expanded Education & Templates Library, aiming to improve board efficiency and risk oversight. [Business Wire]
HelloBoss, a Japanese AI recruitment service, launched its AI Mentor feature offering real-time voice-mock interviews and comprehensive career support for job seekers. [EIN Presswire]
Moveworks has integrated with Box to enhance enterprise search and automation, enabling employees to access and analyze content in Box using AI-powered tools within collaboration platforms like Teams and Slack. [Business Wire]
New Level Work launched Leora AI’s Skills Training mode, offering scalable, personalized leadership development through AI, combining coaching with training classes. [Business Wire]
Nuclearn and the Tennessee Valley Authority have deployed an AI solution to streamline nuclear licensing, enhancing compliance efficiency. [PR Newswire]
Permutable AI has launched a generative AI-powered API for commodities trading, offering real-time insights and advanced analytics to institutional traders. [EIN Presswire]
The Twos app uses AI for ‘smart suggestions’ to enhance productivity by helping users take the first step in their tasks. [The Verge]
Risks
Google's AI Overview confidently presents the fake ‘Kyloren syndrome’ as a real medical condition. [The Decoder]
Columbia’s Tow Center found ChatGPT’s search tool frequently misidentified article quotes, returning false responses on 153 occasions. [The Verge]
Cybersecurity experts warn that generative AI is enabling hackers to easily create sophisticated macOS malware, challenging macOS’s perceived security advantage over Windows. [TechRadar]
LLM-powered robots are vulnerable to hacking, allowing for harmful actions in the physical world, underscoring risks of using AI in safety-critical applications. [Wired]
A Stanford professor has apologized for submitting AI-generated fake citations in a legal defense of Minnesota’s unconstitutional deepfake law. [TechDirt]
Responses
The Anthropic Fellows Program is a six-month initiative to support and fund 10-15 AI safety researchers,. [Anthropic]
AWS launched Automated Reasoning checks to combat AI hallucinations by validating model responses against customer-supplied data, part of AWS Bedrock’s new features. [TechCrunch]
The US Department of Defense has contracted Hive AI for US$2.4m to enhance deepfake detection to bolster national security against AI-driven disinformation and fraud. [MIT Technology Review]
The LA Times plans to introduce an AI-supported ‘bias meter’ to help readers discern news versus opinion. [The Wrap]
Meta reported that AI-generated content constituted less than 1% of election misinformation, countering fears of significant AI-driven disinformation impacts. [Engadget]
MLCommons has launched AILuminate, an industry-standard safety benchmark for assessing LLMs, collaboratively designed by AI researchers and industry experts to improve AI safety transparency and reliability. [Business Wire]
OpenAI’s ChatGPT utilizes content filters to block specific names, like Brian Hood and Jonathan Turley, to avoid generating problematic outputs. [Ars Technica]
Regulation
An Australian Senate Select Committee’s report criticizes big tech companies’ lack of transparency and labels AI models as high-risk, recommending regulatory measures to protect workers’ rights and address AI’s workplace impacts. [TechRepublic]
And Australia’s competition watchdog has urged the revisiting of efforts to ensure more internet choices due to concerns over Google's dominance and AI integration’s potential impact on search accuracy. [Yahoo Finance]
EU antitrust regulators are investigating Nvidia's US$700m bid for Run:ai, focusing on potential discounts offered for bundled GPU software and hardware. [Yahoo Finance]
Under Trump, US AI policy may shift with potential dismantling of Biden’s executive order on AI, affecting innovation, regulation, and related issues. [MIT Technology Review]
The US FTC is investigating Microsoft’s partnership with OpenAI. [Yahoo Finance]
Conversational AI
AKOOL and LiveX AI have partnered to enhance customer interactions by combining advanced conversational AI with dynamic, adaptive avatars for improved engagement and support. [Business Wire]
AMC Technology's DaVinci AI Starter Kit enables businesses to integrate Microsoft Copilot for improved customer interactions and operational efficiency in contact centers. [PRWeb]
The Chronicle of Higher Education launched Chron, an AI chatbot using its archive to address higher education queries. [Business Wire]
Forethought's Agentic AI enhances customer experience by integrating AI-driven automation with human-centric support, achieving over one billion monthly interactions and earning recognition for innovative customer service solutions. [Business Wire]
Kustomer introduced Tasks and Skills-Based Routing to enhance workflow management and personalize customer support. [Business Wire]
Great Southern Bank leverages NICE's CXone Mpower platform to enhance customer service efficiency through AI-driven automation, significantly reducing customer wait times. [Business Wire]
RingCentral has launched Zendesk Talk Partner Edition for RingEX, enhancing Zendesk integration with advanced call handling, data management, and AI-driven insights for improved customer interactions. [Business Wire]
SoundHound AI and Torchy’s Tacos have launched a voice AI Smart Ordering system across all 130 Torchy’s locations. [Business Wire]
Verint has introduced the CX/EX Scoring Bot for real-time customer and employee experience insights, aiming to enhance operational efficiency in contact centers. [Business Wire]
Voice News
ElevenLabs’ GenFM feature converts documents into personalized podcasts with AI hosts discussing the content in 32 languages. [TechRadar]
HitPaw launched VoicePea V2.3.0, featuring a new text-to-speech tool for easily converting text into high-quality, natural-sounding speech, enhancing creative expression. [PR Newswire]
Hiya's Deepfake Voice Detector, a free Google Chrome plugin, uses AI to accurately identify manipulated audio within one second, assisting with real-time deepfake detection. [TechRadar]
Hume AI's Voice Control tool lets users create custom AI voices by adjusting vocal characteristics without any coding, enhancing emotional intelligence and avoiding voice cloning. [VentureBeat]
Document AI
Accelario, a provider of data provisioning and compliance solutions, announced the launch of its Automatic Recognition of Sensitive Data feature. [Business Wire]
Amazon’s newly released second-generation Kindle Scribe includes an AI summarization tool. [GeekWire]
SquareWorks Consulting has launched Invoice AI, a technology that improves invoice scanning accuracy and efficiency, surpassing traditional OCR solutions. [Business Wire]
Other Tongues
Indigenous in AI, an international community of Native, Aboriginal and First Nations engineers, is leveraging AI to preserve endangered languages and cultures. [NBC News]
Lilt unveiled new AI features for enterprise translation and content creation, enhancing AI management, multilingual results, and video translation efficiency. [PR Newswire]
MotionPoint has introduced Adaptive Quality Estimation, a cutting-edge technology designed to automate translation quality estimation and streamline the quality control process. [MultiLingual]
A Japanese publisher, Orange, uses AI to translate manga, but this approach has sparked controversy among some fans over its impact on traditional artistry. [MIT Technology Review]
Trados announced Generative Subtitles, an AI-powered video subtitling feature, alongside several platform enhancements, including improved translation tools. [Slator]
Search
Algolia’s AI Search and Discovery platform is now available with Pay-As-You-Go pricing in AWS Marketplace. [Business Wire]
Apple endorses AWS by using its AI chips for search services. [CNBC]
The Browser Company is launching Dia, an AI-powered web browser designed to simplify internet tasks. [The Verge]
Cohere's new search model, Rerank 3.5, aims to revolutionize global business data retrieval with multilingual capabilities and enhanced AI reasoning. [VentureBeat]
Surf, a new AI-powered browser by Deta, uses built-in chat for semantic search, context organization, and web interaction. [The Verge]
Dropbox's Dash for Business centralizes AI-powered workplace searches to improve productivity. [TechRadar]
Berlin-based Ecosia and Paris-based Qwant, two of Google’s smaller rivals, are launching their own indexing project. [Fast Company]
Elastic has launched Elastic Cloud Serverless, a scalable, AI-optimized search platform, offering flexible cloud solutions for search, observability, and security applications. [Business Wire]
Encore is an AI-powered platform that facilitates online second-hand shopping using conversational search. [Wired]
Startup Exa aims to convert the chaotic web into a precise database using LLMs, offering detailed search results through its engine, Websets. [MIT Technology Review]
Twelve Labs is using AWS to develop and scale AI models that enhance video understanding, enabling advanced applications like video search and content summarization. [Business Wire]
Health Tech
ClinicMind’s AIScribe automates clinical documentation to enhance workflow efficiency, reduce administrative burdens, and improve patient care quality in healthcare practices. [PRWeb]
Hippocratic AI has partnered with the Nurses on Boards Coalition to enhance nurses’ influence in healthcare technology through education and leadership development initiatives. [Business Wire]
iMerit launched ANCOR, an AI-driven Annotation Copilot for Radiology, enhancing radiology AI development efficiency and accuracy. [PR Newswire]
Maverick Medical AI's real-time AI-powered coding solution improved RadNet's efficiency by automating medical coding across 399 sites. [Business Wire]
Medsender launched MAIRA, a multilingual AI voice agent, to provide 24/7 support in healthcare, enhancing patient communication and automating practice workflows. [Speech Technology Magazine]
OpenAI’s ChatGPT outperformed doctors in diagnosing medical conditions from case reports. [Futurism]
Healing Hands Ministries in Dallas is using Sunoh.ai's AI medical scribe to enhance clinical documentation efficiency. [Business Wire]
Legal Tech
Assembly Software announced NeosAI, an AI-powered enhancement to Neos, offering law firms increased efficiency, reduced errors, and streamlined case management with customizable features. [GlobeNewswire]
Automatise’s Cicero, powered by Meta's Llama models, enhances legal document analysis by providing five core document-related functions. [Artificial Lawyer]
Consilio announced the launch of Native AI Review, a private cloud solution integrating AI into third-party legal review tools, enhancing eDiscovery with tailored LLM innovations. [Business Wire]
ContractPodAi announced Leah Drive, an AI-powered hub for managing and analyzing legal documents. [Business Wire]
Curvestone launched Workflows for its WorkflowGPT platform, enabling law firms to create custom AI processes for efficient legal document analysis and service agreement reviews. [Artificial Lawyer]
Integra Ledger has launched ‘Integra Ledger Connect’ to enhance document authenticity and trust in contracts using blockchain technology, particularly against generative AI-related risks. [Artificial Lawyer]
Law Practice AI has introduced Demands AI and Doc Reader AI to enhance efficiency, accuracy, and compliance for legal and medical professionals. [GlobeNewswire]
Linklaters has launched an AI Sandbox to develop generative AI solutions from staff ideas. [Artificial Lawyer]
The LITIG AI Benchmarking project is forming a working group to draft a consultation paper on generative AI accuracy. [Artificial Lawyer]
Noxtua, a European legal AI developed by Xayn, is partnering with ClauseBase to enhance legal document automation with privacy-focused LLM capabilities. [Artificial Lawyer]
Panoram launched PanoramAI, a generative AI document review platform using multiple models to enhance speed and accuracy. [Legal IT Insider]
Spellbook has grown to serve 2,600 customers globally with an AI suite for transactional lawyers, featuring new branding and five distinct features. [Artificial Lawyer]
Springbok AI has launched SpringLaw Verify, a tool for detecting hallucinations in legal documents, enhancing accuracy and trust in AI-generated content. [Legal IT Insider]
Funding
9fin, an AI-driven debt capital markets platform favored by top law firms, raised US$50m in Series B funding. [Artificial Lawyer]
Bounce Insights, an Irish startup specializing in AI-powered market research, secured €4.2m in funding to expand into the US. [EU-Startups]
Cake, an AI infrastructure platform, raised US$13mm to make cutting-edge AI accessible to mid-market companies. [Business Wire]
Cresta raised US$125m in Series D funding to enhance its AI platform for contact centers. [FinSMEs]
Enterpret, an AI-enabled customer feedback platform, raised US$20.8m in Series A funding to scale operations and deploy no-code AI agents for product development and CX teams. [FinSMEs]
HappyRobot, a San Francisco-based AI voice solutions company for logistics, secured US$15.6m in Series A funding to enhance its operations and development efforts. [FinSMEs]
Lawhive raised US$40m to expand its affordable legal services in the US, leveraging AI to automate tasks and offering fixed fees to increase accessibility. [Artificial Lawyer]
Lica raised US$4m for its AI tool that simplifies video creation from screen recordings, targeting prosumers and business teams. [TechCrunch]
Nvidia has joined a US$700m Nebius deal for AI cloud services. [Bloomberg]
Pocketlaw raised nearly €5m in funding to accelerate AI development and European expansion. [Artificial Lawyer]
Recall, an Amsterdam-based AI-powered personal encyclopedia, raised US$1.5m to expand operations and offers features like summarizing, categorizing, and connecting content in a personalized knowledge graph. [FinSMEs]
Roon, a health information platform, raised US$15m to expand its medically-vetted resources, focusing on women’s health, cancer care, and more. [Tech Funding News]
Tenstorrent raised US$693m in Series D funding to expand its engineering team and develop AI training servers, valuing the company at US$2.6 billion. [TechCrunch]
London-based startup Vente AI raised £500K to revolutionize recruitment with AI, aiming to enhance efficiency and competitiveness by automating lead identification and prioritization. [Tech Funding News]
London-based Vinter raised £1.1M Seed funding to expand its AI recruitment platform that modernizes workflows and promotes fair hiring. [FinSMEs]
Wexler AI, a startup enhancing legal fact-checking and intelligence in high-stakes disputes, secured US$1.4m to expand its offerings and launch in North America. [Artificial Lawyer]
Yurts, which deploys AI in high-security sectors, raised US$40m to support contracts with the US DoD. [TechCrunch]
Acquisitions
Captions, an AI video creation startup, acquired AlpacaML after raising US$60m in Series C funding. [Slator]
Imagine Learning has acquired Pango Education to enhance its AI-driven curriculum solutions, aiming to improve personalized learning and support educators in creating effective lesson plans. [Business Wire]
Uniphore is acquiring ActionIQ and Infoworks to expand its Enterprise AI platform with a Zero Data AI Cloud, enabling seamless AI adoption without complex data integration. [Business Wire]
Uptiq.AI has acquired UpSwot to enhance its AI platform for financial services, integrating UpSwot’s data gateway to improve AI-driven applications and deliver innovative solutions. [Business Wire]
There’s More
Tim Cook discusses Apple's strategic entry into generative AI with Apple Intelligence, emphasizing privacy and long-term, highly considered innovation in a competitive tech landscape. [Wired]
Critics argue that startup Spines, which charges authors to publish books using AI, prioritizes profit over quality and genuine literary interest. [The Guardian]
The end of AI scaling may not be nigh: Here’s what’s next. [VentureBeat]
GAI Insights has released a buyer’s guide to LLMs, identifying seven emerging leaders in generative AI. [VentureBeat]
OpenAI's partnership with Anduril marks its significant shift towards military applications, enabling AI deployment on battlefields despite previous prohibitions against weapon development. [MIT Technology Review]
The Back Cover
Here’s Winter Youth with a cover of David Bowie’s Changes.
If you know of a great cover version of a back-catalogue classic, drop an email to us at backcover@language-technology.com and we’ll consider it for inclusion here.
Got this from someone else?
Did you find this newsletter useful? if so, you might forward it to a friend; or you could email us at news@language-technology.com to tell us what you want more of.
Did you hate this newsletter? if so, you could forward it to an enemy; or you could email us at news@language-technology.com to tell us why – make it better!