Archive

X.AI has announced a significant update to its Grok AI model, making it available to all users on the X platform (formerly Twitter) at no cost.

Anthropic Reports Minimal Election-Related Usage of AI Assistant Claude in 2024 In the first major election cycle with widespread access to generative AI

Google has unveiled Gemini 2.0, its latest artificial intelligence model designed to understand and act on information across multiple formats.

X.AI has announced the integration of image generation capabilities into its Grok AI system through a new model code-named Aurora

NVIDIA has announced major initiatives to advance sovereign AI development in Southeast Asia, with founder and CEO Jensen Huang meeting with the prime ministers

Perplexity AI has announced a significant expansion of its Publishers' Programme, welcoming over a dozen new media partners from across the globe.

ChatGPT users will soon have access to a wealth of specialised content through a new strategic partnership between OpenAI and Future, the global

Salesforce revenue hit $9.44B (+8% YoY) in Q3 FY25. Stock jumped 9% to $361.23 as Agentforce AI adoption drives growth. Operating margin reaches 20%, up from 17.2%.

Legal fights: Musk seeks injunction on OpenAI-Microsoft deals; Canadian media claims 16.1M works used without permission for AI training

Google's London AI Campus launches with £865K grant to Raspberry Pi Foundation, targeting 250K students by 2026 via Experience AI programme

Claude gains preset/custom communication styles. GitLab uses feature for business cases, docs & marketing content, with style-specific memory

£8.22M UK lab to counter Russia & North Korea's AI threats; NK already using AI to accelerate malware and scan security vulnerabilities

Claude models on AWS Bedrock power 80% faster document search in multiple languages, while Perplexity achieves 2x processing speed with improved accuracy

41% of firms use AI copilots in customer service. NVIDIA's NIM platform enables LLMs, knowledge retrieval & digital human interfaces for agent ops

Q3 revenue hit $35.1B (+94% YoY) amid global AI infrastructure rollouts and sovereign computing projects.

Meta hire ex-Salesforce AI chief to lead new Business AI unit, building on 200M business users and Llama's 600M downloads for enterprise expansion

Free beta offers enterprise features like expense reporting and meeting summaries, challenging paid platforms while raising shadow AI adoption concerns.

Mila's report spans health, genetics, neuroscience, materials discovery, climate research & Indigenous tech integration, with 1,300+ researchers involved.

OpenAI partners with Sanofi, Simplon, Mirakl & ESCP in France, focusing on clinical trials, education, marketplace growth & personalised learning.

NVIDIA leads 750+ open-source projects, contributing to CNCF initiatives over a decade to democratise AI development tools and cloud tech.

GluFormer uses GPT-style transformer architecture to predict glucose from 15-minute readings, foreseeing responses to specific foods and tracking multiple health metrics.

New prompt improver uses chain-of-thought, XML standardisation, and prefill features to enhance prompts, boosting accuracy by 30% in classification tests.

Company ranked first among 2,000 evaluated firms, excelling in employee benefits with quarterly Free Days and onsite counsellors, while maintaining low turnover

Company created 240 custom GPTs in 10 weeks, including specialised tools for fragrance analysis and clinical trials, achieving 90% faster response times.

The initiative runs on GPU Merdeka, an NVIDIA-powered sovereign AI cloud built in three months. It operates from a renewable energy data centre.

User cases show 75% HR containment, 93% ticket accuracy, 50% faster case prep, and 95% reduced editorial review with Llama on watsonx.

Models combining web search with LLMs outperformed standalone models on SimpleQA. Real-time data access proved crucial for factual accuracy.

CAISI launches with $50M budget, linking NRC, CIFAR & 3 national AI institutes. Will research AI safety, focusing on disinformation and cybersecurity threats

AI search engine introduces sponsored questions & paid placements while preserving unbiased results. Revenue aims to support publisher partnerships

Salesforce advances AI integration through Agentforce platform, combining AI agents with humans to enhance customer relationship management

System achieved 7-day accuracy matching old 5-day forecasts via tripled training sites, rebuilt architecture, and enhanced weather data integration.

LLMs show high consistency on neutral topics like Thanksgiving but become variable on controversial issues. Larger models outperform smaller ones in reliability.

What seemed an expensive Twitter purchase evolved as xAI's valuation reaches $45B, while X itself transforms from social media platform to AI data engine.

CurePharma AI developed a system using Llama models to scan prescriptions, translate medical info to local languages, and connect users to e-pharmacies

The partnership combines DiaGen's AI drug discovery with Mila's ML research to optimize therapeutic proteins for stability and target binding.

Ochoa's appointment expands the board to 13 members, with CEO Huang citing her trailblazing leadership in space exploration and governance roles.

Users can process high-volume requests at half the cost of regular API calls, with applications in sentiment analysis, translation, and vector embedding.

The method uses machine learning to compute mathematical variances revealing gene influences, analyzing relationships layer by layer to identify regulatory groups.

While software security has reporting infrastructure and bug bounties, AI systems lack similar frameworks for third-party evaluation and protection

The AI system integrates camera footage, weather data, and construction updates to detect anomalies, improving incident response time by 8.5 minutes.

"As election results loom, xAI's $15.4B in government contracts and stance on AI bias reflects deepening ideological divisions in Silicon Valley's AI landscape."

"AI radio hosts replaced staff but failed after airing synthetic interview with deceased poet. Tech worked, but human judgement sparked public backlash."

"Legacy systems make up 50% of app estates. Banks face $100M+ modernisation costs for AI readiness, while retailers see ML benefits in supply chains."

"MIT tests show LLMs fail at NYC navigation when 1% of streets close. Internal maps reveal imagined streets and impossible layouts despite high performance."

"MIT's ICSR Data Hub stores police data from 40 major US cities, using AWS to study systemic racism's patterns through machine learning analysis."

"New Grok-beta model features 128k token context and function calling, with REST API compatibility with OpenAI/Anthropic. Multi-modal version coming soon."

"AI agents analyse video/image content using vision language models, allowing natural language customisation instead of coding for visual monitoring and alerts."

The AI lipsyncing app achieved 20M downloads and 100M created clips in 3 weeks. Their AI art generator Wombo Dream later won Google Play's Best App 2022.

"Perplexity's new AI tool offers live election results, voting info, and candidate positions, partnering with AP for real-time updates."

"The search uses fine-tuned GPT-4 with novel synthetic data generation. Users can trigger web searches automatically or manually, with linked source citations."

Talking Tours' uses AI for real-time audio commentary at cultural sites, while 'Mice in the Museum' features AI characters guiding through 100+ collections.

"Meta's new robotic fingertips can feel the lightest touch. Their testing system puts robots through everyday tasks in virtual homes with thousands of items."

"AI turns black-and-white MRI scans into color 3D models, showing tumors in blue and veins in red, helping surgeons plan better cancer operations."

"AI systems went from solving 2% of real coding problems to 49% in one year. Anthropic warns the window for safe regulation is closing fast."

"Stanford's fellowship places tech experts in government roles. One fellow helped write AI laws, while others improved public services and training."

OpenAI's SimpleQA tests 4,326 factual questions with 3% error rate. GPT-4o scores under 40%, showing larger models excel while deeper thinking ones opt to decline.

New MIT postdoc programme pairs AI experts with mentors in biology, physics, music & more, funded by $20M Tayebati gift to advance cross-disciplinary AI applications.

Google's Trillium TPU achieves 4.7x peak compute vs prior gen with 67% better energy efficiency, now powering Search, YouTube & DeepMind's LLMs in preview.

Using Edit & Bash tools, newest Claude 3.5 Sonnet hits 49% on SWE-bench's GitHub issue tests, despite challenges like hidden tests & high token costs.

NVIDIA's Q3 fiscal 2025 results to be discussed Nov 20, with CFO Colette Kress releasing commentary prior. Q&A limited to analysts and institutional investors.

"ChatGPT-4 scored 92 in clinical reasoning vs physicians' 74-76. AI-assisted doctors completed diagnoses 1+ minute faster but showed no accuracy gains"

"Agentforce enables autonomous AI agents triggered by data/rules, built via no-code tools, integrated with Data Cloud/MuleSoft/Slack for 24/7 automation"

"Decagon combines GPT-3.5 for query rewriting in RAG workflows with GPT-4 for complex decisions, achieving 91% automation in customer support"

"Claude 3.5 Sonnet scored 93.7% on HumanEval, tops SWE-bench Verified, offers real-time debugging and test generation via Amazon Bedrock inference"

"Enterprise RAs combine NVIDIA-Certified servers, Spectrum-X AI Ethernet, BlueField-3 DPUs with NeMo and NIM microservices for optimized AI infrastructure"

"Using H100 GPUs & Llama 3, Ntropy achieves 20x model efficiency, while NayaOne's AI sandbox delivers 10x faster fraud detection with 40% cost reduction."

"Voice tech update adds 12 African languages to Search & Gboard, with $1.5M to train Nigerians in AI/data skills via Data Scientists Network Foundation."

"New AI system shows 40% better fraud detection, uses RAPIDS, GNN & XGBoost. Bunq reports 100x faster model training; AmEx makes fraud calls in milliseconds."

"While AI tools save time, human oversight remains essential. As AI-360 notes: 'HUMAN IN THE LOOP is very much ESSENTIAL in largely, every step.'"

"AI transformed Brenda Lee's 'Rockin' Around The Christmas Tree' to Spanish with artist consent, producer Baqueiro and vocalist Hoyle collaborating."

"Meta partners with Reuters while NYT litigates, showing data access divide: only AI companies with 'deep pockets' can afford premium content deals."

"HPT unifies 200,000+ robot trajectories across 52 datasets, processing vision and proprioception inputs equally, improving performance by 20% in real tests"

"Studies found fabrications in 80% of public meetings, 50% of 100+ hours of audio, and nearly all 26,000 transcripts analyzed by developers"

"Next year's AI surge may displace 15-20% of jobs. Beyond market values, success hinges on managing workforce transformation and societal impact."

"Universities split on AI policy: some ban it outright, others allow grammar checking, while detection tools struggle with 68-84% accuracy rates."

"Research spans autonomous surgery systems, smart assistants for medical diagnosis/climate prediction, and AI-art integration for creative interpretation."

"New code sandbox allows Claude to run JavaScript, clean data, create visualisations, and perform real-time analysis on uploaded CSV files."

"NVIDIA plans 10x GPU expansion in India, partnering with tech firms to train 500k developers in sovereign AI, agentic AI, and physical AI systems."

"Quantised Llama 3.2 achieves 56% size reduction using QLoRA and SpinQuant, with 4-bit weights and 8-bit activations for mobile deployment."

"SAM 2 enables real-time object segmentation, processing 60M+ polygons across 350M images for manufacturing QA and scientific research."

"AI search engine defends right to cite public facts, highlights source attribution system and revenue-sharing with TIME, Fortune, Der Spiegel."

sCM system reduces image generation to 0.11 seconds using two processing steps vs hundreds, and scales to 1.5B parameters while matching quality of slower methods.

Gefion supercomputer, with 1,528 H100 GPUs, enables quantum simulation jump from 36 to 40 qubits while accelerating weather forecasts from hours to minutes

Meta's Llama model uses split inference: 1st layer processes on user devices, remaining 31 in cloud, enabling privacy while avoiding quantisation

Google Cloud's Vertex AI powers GOV.UK's new search, showing 14% higher click-through rates and reduced multiple searches for public services

MusicFX DJ adds 60-second sharing, speed/brightness controls, and continuous music flow, developed with Jacob Collier and Google DeepMind

''Plug and play AI now moves cursors, clicks, and types through virtual keyboards, moving beyond traditional AI-specific interface requirements."

"Mila and SonoNurse partner to develop AI that validates medications pre-administration, combining deep learning with medical device expertise."

"29% of UK doctors use AI, but 70% lack AI training. Medicine and radiology show highest adoption rates, with most doctors seeking better preparation."

"SMAST uses selective attention and motion tracking to analyse human actions in video, outperforming existing systems in benchmark testing."

"Participants showed protective behaviour toward excluded AI bots during virtual games, with older players particularly sensitive to AI ostracism."

"Google's AI education programme piloted in 8 schools before reaching 22,000 UK students, teaching AI identification and critical assessment skills"

"Only 28 of 50 federal agencies published AI compliance plans, with none submitting impact assessments or extension requests for non-compliant systems"

"NVIDIA NIM, ACE, and Riva tech powers multilingual AI avatars at Ottawa Hospital, offering 24/7 pre-surgical guidance through natural conversations"

"NVIDIA expands Indian operations across 6 locations with 4,000+ employees, supporting 2,000 startups via Inception programme and deploying Omniverse/Isaac tech"

AI system processes 32 satellite frames per location using teacher-student model to super-resolve low-res images and extract building footprints with 1.5m accuracy

"The x86 Ecosystem Advisory Group brings together former rivals Intel and AMD, with tech leaders like Torvalds, Microsoft, Google, and Meta joining forces."

"OpenAI's defensive patent pledge faces skepticism as they block competitor access to training data while defending against copyright infringement suits."

X-series chips use squeezers, interferometer networks, and photon detectors for room-temperature quantum computing, targeting million-qubit scaling.

Paris-based Zelros transforms insurance with AI-powered SaaS. Founded 2015, raised $16M+. Their platform streamlines sales and claims with real-time recommendations.

Stanford's RegLab uses AI to identify racial covenants in 5M+ property deeds. The system saves 86,500 person-hours, costs 2% of proprietary models, reveals patterns.

Meta FAIR releases SAM 2.1 for improved image segmentation, Spirit LM for multimodal speech/text, Layer Skip for faster LLM inference, and tools for cryptography.

PRH prohibits book use for AI training, aims to protect authors' IP. Industry bodies welcome move. Legal experts note complexities in copyright issues for AI.

NVIDIA and Siemens use digital twins to boost sustainable manufacturing. Foxconn's virtual plant on Omniverse aims to cut energy use by 30%. GPU-CPU systems are 20x more efficient for AI tasks.

"Apple Intelligence rollout begins with notification summaries, expanding features through March 2025 as all screen devices become AI-capable by 2026."

Meta's AI video tool, Movie Gen, is being tested with Blumhouse and select filmmakers. It generates custom videos from text, edits existing ones, and transforms images into videos.

Clearbot's AI-driven boats use NVIDIA Jetson tech to autonomously clean waterways. They collect tonnes of waste monthly in India, with geotagged data aiding pollution tracing.

Study shows US clinicians lose 28 hours weekly to admin tasks. 90% of providers optimistic about AI's potential to reduce paperwork, freeing time for patient care and reducing burnout.

Perplexity's AI platform now offers Internal Knowledge Search, combining web and internal data search. New Spaces feature enables AI-powered team collaboration on research projects.

NotebookLM now offers customisable Audio Overviews and background listening. A new Business version is coming, tailored for organisations. Over 80,000 companies already use it.

The Alan Turing Institute initiates a redundancy process, potentially cutting jobs from its 440-strong workforce. The strategy shift aims to focus on health, the environment and defence, citing reduced core funding.

Mila and Ability Biologics combine AI and immunology expertise to develop logic-gated antibodies responding to cellular microenvironments, accelerating therapeutics.

Ministral 3B and 8B models outperform larger peers, support 128k context, and enable on-device AI for robotics and local analytics. Pricing from $0.04/million tokens.

CodeGPT's Llama integration boosted productivity 30%, sped up onboarding, and enabled codebase graph understanding. Open-source nature crucial for innovation.

Prophesee is a French startup that has made significant strides in the field of bio-inspired computer vision systems since its founding in 2014.

Google Shopping overhaul uses Gemini AI models with 45 Billion product listings. Features: AI-generated briefs, personalised recommendations, virtual try-on.

ChatGPT showed minimal bias across gender, race, ethnicity. LMRA analysed millions of interactions. <0.1% cases had harmful stereotypes. Newer models improved.

Lenovo & NVIDIA unveil Hybrid AI Advantage platform. AI Fast Start service for rapid deployment.

Anthropic updates Responsible Scaling Policy. New capability thresholds for AI safety. ASL Standards introduced. Kaplan as Responsible Scaling Officer.

Google's deal with Kairos Power for SMR nuclear energy aims to meet growing electricity demands driven by AI technologies powering scientific advances.

AI adoption widespread in law firms, used for back-office tasks and increasingly for client-facing work. Larger firms lead in sophisticated integration.

Stanford study: LLMs can assess and optimise educational materials, replicating learning effects and generating content preferred by human teachers.

King's College engineers create fluid-based circuits to control robots without electricity, freeing up 'brain' space for advanced thinking.

US considers country-specific limits on AI chip exports, focusing on Persian Gulf. Follows recent rule easing shipments to Middle East data centres.

Wimbledon replaces human line judges with AI from 2025. Meanwhile, Klarna and ByteDance cut jobs, leveraging AI for efficiency. This highlights accelerating AI adoption and growing job displacement concerns across industries.

"Harvard students showed Meta Ray-Bans 2 can be linked to facial recognition and people-search databases to identify strangers instantly via I-XRAY system."

Microsoft urges holistic modernisation for AI readiness. Successful organisations build clear strategies around AI and AI training, tying them closely to business outcomes.

NVIDIA's AI Summit highlighted AI's role in AV safety. MITRE and Mcity partnered to develop a virtual/physical AV validation platform using NVIDIA tech.

Microsoft's new medical AI models for image analysis and report generation.

Lisbon's Unbabel combines AI and human expertise for business translations. Founded in 2013, raised $91M. Serves major clients like Booking.com, in upto 28 languages.

Lucy AI model generates custom audio ads with dynamic localization, incorporating real-time data on pricing and conditions for retail networks.

Energy Secretary Granholm: AI crucial for global challenges. DOE focuses on responsible AI use. Data centres to increase power demand by 15% on US grid.

Therapyside uses NVIDIA AI for virtual assistants in therapy. Platform transcribes sessions, summarises notes, and answers queries via RAG technology.

OpenAI's MLE-bench tests AI on ML tasks. Best AI setup achieved bronze-level in 16.9% of competitions. Study explores resource-scaling and contamination effects.

Google supports startups using AI for mental health. O7 Therapy matches patients with therapists, Callyope analyses speech, HearMe provides AI chatbots, and Thalia assists therapists in analysis.

SETI Institute uses NVIDIA AI to detect faint radio signals from space in real-time. Pilot project processed 90 billion data packets at 100Gbps in 15 hours.

NVIDIA's NIM Agent Blueprint for container security uses AI to reduce vulnerability analysis from days to seconds. Deloitte adopts it for enterprise solutions.

NVIDIA partners with tech giants to develop custom AI apps using NIM and NeMo tools. New microservices streamline generative AI development for various industries.

NVIDIA's AI tools accelerate healthcare innovation. NIM microservices and Agent Blueprints aid in medical imaging, drug discovery, and data extraction from PDFs.

71 PhD students join Turing Institute's Enrichment scheme, enhancing AI research. New Community Award enables 20 students to engage remotely, expanding collaboration.

Universities collaborate on DARPA-funded AI tools for turbomachinery design. The project aims to optimise shape and material composition for aerospace components.

Neuromnia uses Llama 3.1 to create Nia, an AI co-pilot for ABA therapy. Fine-tuned with expert data, it automates treatment planning and documentation for autism care.

NVIDIA optimised LLMs, achieving 3.5x latency improvement for Llama 70B in under a year. Blackwell platform shows 4x performance boost and first FP4 precision use in MLPerf.

OpenAI partners with Hearst to integrate content from 20+ magazine brands and 40+ newspapers into ChatGPT, enhancing information base.

MITRE and Mcity partner to create a virtual and physical AV validation platform, addressing regulatory challenges with comprehensive resources open to the industry.

NVIDIA, Moderna, and Yale review how quantum machine learning may enhance drug discovery. GPU-accelerated simulation of quantum algorithms is key to this research.

NVIDIA launches AI Tech Community in Pittsburgh, partnering with CMU and Pitt to establish joint centers for robotics, AI, and intelligent systems research.

TSMC adopts NVIDIA's cuLitho platform for chip manufacturing, using GPU acceleration and generative AI to speed up computational lithography by 2x.

Anthropic implements measures to prevent AI misuse in 2024 US elections, including policy updates, detection systems, and redirects to voting information.

Anthropic launches Message Batches API, allowing processing of up to 10,000 queries within 24 hours at half the cost of standard API calls.

Google Cloud expands UK data residency for Gemini 1.5 Flash AI, allowing organisations to process machine learning data within UK borders.

This supercomputer is poised to accelerate breakthroughs in fields ranging from healthcare to urban planning, potentially reshaping industries and improving lives across the island and beyond.

Movie Gen combines multiple AI models for video, image, and audio generation. It can create 16-second videos from text, edit existing videos, and generate 45-second audio clips.

Foxconn's cloud and networking products division saw robust revenue growth due to strong AI server demand.

AI system processes 500M monthly visual searches, using GenAI tagging and multi-search tools to enable hybrid product discovery across platforms.

Berlin's Acrolinx leads AI-powered enterprise content governance. Founded in 2002, raised $60M. Serves Fortune 2000 with real-time content optimisation.

As OpenAI secures record-breaking funding and seeks investor exclusivity, questions arise about the future of AI development, ethics, and industry competition. In a move that has sent ripples through the tech world, OpenAI, the creator of ChatGPT, has announced a staggering $6.6 billion funding round, valuing the company at

Art Transfer 2 uses AI to transform any photo into various artistic styles, providing insights on brushwork, colour, and composition. Culture Weekly offers personalised content.

OpenAI secures $6.6B in funding and $4B credit facility, totaling $10B+ in liquidity to fuel growth, research, infrastructure expansion, and talent acquisition.

Bioeureka's AI-based app uses smartphones to quickly detect pathogens, addressing the need for rapid, reliable testing in healthcare and food industry sectors.

OpenAI's Canvas offers inline AI feedback, shortcuts for tasks like adjusting text length and debugging code, and allows users to highlight sections for focused AI assistance.

Google Lens now offers video understanding, allowing users to ask questions about moving objects in real-time. AI processes video and queries to generate informative overviews.

AI systems on accelerated computing platforms have become 45,000 times more energy-efficient in 8 years. GPUs and CPUs combined are 20x more efficient than CPUs alone.

MIT team uses AI to improve simulation accuracy. New MPMC method with graph neural networks enhances data distribution in complex spaces.

Microsoft relaunches AI 'Recall' tool with privacy upgrades. Now opt-in, biometric login required. Exclusive to new CoPilot+ AI-powered laptops.

Author sues Meta for using pirated books to train Llama AI. Lawsuit claims copyright infringement, highlighting tension between AI and IP rights.

OpenAI raises $6.6B at $157B valuation. ChatGPT, their AI chatbot, now has 250 million weekly users worldwide, showing rapid AI adoption.

MetaLearner integrates Llama 3.1 models, enhancing enterprise analytics. Open-source AI enables rapid innovation in data science.

Material Legal Technologies Inc partners with Mila to develop AI solutions for financial and legal document processing.

Court ruled LAION's use of images for AI dataset was legal under German law for scientific research. Case may impact future AI copyright disputes.

Oracle's $6.5B investment brings advanced AI infrastructure to Malaysia, offering 150+ solutions including AI and data analytics tools to boost competitiveness.

OpenAI introduces vision fine-tuning for GPT-4o, Realtime API for multimodal experiences, and Model Distillation suite for cost-efficient AI development.

DKFZ's TrainBot, a convoy of miniature robots, performs endoscopic microsurgery. Wirelessly controlled, it can transport heavy instruments to hard-to-reach areas.

NVIDIA's AI Summit to showcase AI in healthcare, cybersecurity, and manufacturing. 50 sessions cover generative AI, remote sensing, and industrial digitalisation.

Malaysia aims to lead ASEAN in AI, partnering with Google. PM Anwar launched AI initiatives, including talent roadmap and consortium, to boost development.

Generative AI could add €1.2-€1.4 trillion to EU GDP in 10 years. 61% of jobs may be augmented by AI, with only 7% facing long-term automation.

Tokyo-based Preferred Networks pioneers deep learning for robotics & edge computing. Valued at $2B, they partner with Toyota & FANUC for industrial AI solutions.

MIT's 'Future You' AI system enables users to converse with simulated 60-year-old selves, aiming to reduce anxiety and improve long-term decision-making.

MIT create a dataset of 5,000 'illusory' faces, revealing differences between human and AI face detection and potential links to animal face recognition.

MIT's Clio enables robots to map scenes and identify task-relevant objects using natural language prompts, combining computer vision, language models, and mapping tools.

Meta's Reality Labs release a dataset of 2,400+ detailed 3D object models, aiming to accelerate 3D reconstruction research for e-commerce and immersive reality.

Newsom vetoed AI safety bill, citing innovation concerns. The bill would've required kill switches for AI systems and established oversight for advanced models.

MIT researchers use AI to estimate particle size 60x faster in pharma manufacturing, reducing analysis time from 15 to 0.25 seconds.

Mercado Libre launches Verdi, an AI platform using GPT-4o, handling 10% of customer service and potentially supporting 9,000 human operators.

Manhattan Declaration proposes 10 principles for global AI governance. Experts emphasise cooperation, risk assessment, and responsible development.

Llama 3.2 introduces vision models and lightweight text models for edge devices. New 128K token context length enables advanced on-device AI tasks.

ChatGPT users gain access to Italian content from GEDI publications. Partnership aims to improve relevance and accuracy for Italian users beyond ChatGPT.

Microsoft enhances AI security with new features: proactive risk assessments, real-time hallucination correction, and confidential inferencing.

Stanford's "Digitalist Papers" explore AI's impact on democracy, featuring 19 experts' essays on governance strategies in the AI era.

Researchers propose labelling AI systems in healthcare with usage info, including training time, data origin, intended use, biases, and warnings.

Meta awards $1.8M in grants for AI projects using Llama models, supporting education, agriculture, and healthcare initiatives globally.

Wellcome Sanger Institute adopts NVIDIA tech for genomic analysis, reducing runtime by 1.6x, costs by 24x, and energy use by up to 42x.

OpenAI launches Academy programme, investing $1M in API credits to support AI developers in low- and middle-income countries globally.

Google commits $120M for global AI education, expands UN partnerships for SDGs, and launches AI-powered datasets for urban planning.

Hitachi Rail integrates NVIDIA AI for real-time railway monitoring, reducing maintenance costs by 7x and service delays by up to 20%.

Google Cloud showcases 185 generative AI use cases across industries, from customer service to cybersecurity, highlighting rapid enterprise adoption.

NVIDIA joins PGIAI, pledging $10M in free AI training for universities and developers in emerging economies, aiming to foster sustainable development.

NVIDIA launches AI Aerial platform, integrating AI with radio networks. Aims to optimise telecom performance and enable new AI services.

Arco Educação's AI Teacher Assistant uses GPT-4 to create personalised lesson plans, achieving 90% accuracy in Portuguese content generation.

NVIDIA's AI tech drives energy efficiency: Foxconn's digital twin saves 30% energy, cloud services see 33x speedup, genome sequencing 42x more efficient.

Medical experts use NVIDIA-powered federated learning to enhance AI models for tumour segmentation, exchanging models instead of data across institutions.

Genmab's "AI Everywhere" initiative gives 2,000+ employees ChatGPT Enterprise access. Custom GPTs assist in cancer research and clinical trial document creation.

Google Translate added 110 languages via AI. AlphaFold aids 2M+ researchers. AI systems like Flood Hub and FireSat provide early disaster warnings globally.

MIT study: AI models in home surveillance inconsistently detect crime, recommend police calls, and show demographic biases in decision-making

MIT's Entrepreneurship JetPack uses AI to rapidly answer startup queries, streamlining development. It's trained on a 24-step framework and provides instant research.

MIT's Crystalyze AI reveals crystal structures from powder samples. It predicts lattice size, composition, and atom arrangement with 67% accuracy on new data.

NVIDIA's AI Aerial platform optimises wireless networks, enables gen AI on single infrastructure. Collaborates with T-Mobile, Ericsson, Nokia for AI-RAN.

Anthropic's "Contextual Retrieval" RAG method reduces failed retrievals by 49-67%. Preserves context in knowledge bases via Contextual Embeddings and BM25.

Google proposes UK National Research Cloud for AI development and National Skills Service for AI workforce training, projecting £400B economic value by 2030.

Google implements C2PA tech for AI content transparency. "About this image" in Search will show AI metadata. Ad systems to use C2PA for policy enforcement.

Google.org's $25M initiative equips 500,000 educators and students with AI skills. 5 organisations focus on different aspects of AI education integration.

Together AI's LlamaCoder uses Llama 3.1 405B to generate full apps from prompts. 200,000+ apps created in a month, democratising AI development.

Alan Turing Institute: AI disinformation didn't affect UK/EU elections. Few viral cases found, but concerns remain about voter confusion and deepfakes.

HEXEL modules combine soft and rigid components, enabling rapid reconfiguration. Magnets allow quick connections, adapting robots for various tasks and environments.

Mistral AI offers free tier, cuts prices up to 80%. Upgraded 22B-parameter Small model improves reasoning. Pixtral 12B added for free image analysis.

Lowe's uses fine-tuned GPT-3.5 to improve e-commerce data quality. 60% increase in error detection, 20% boost in accuracy for product searches.

Google's FireSat uses AI to detect 5x5m fires globally every 20 mins. Custom infrared sensors and AI compare imagery to previous data for rapid detection.

Salesforce launches Agentforce Partner Network, aiming for a billion AI agents. Partners develop actions and templates for cross-system AI integration.

Salesforce and Google Cloud partner to create AI agents working across Customer 360 and Google Workspace, enhancing cross-platform productivity.

Salesforce and NVIDIA partner to enhance AI agents and avatars. Aim to accelerate AI development, improve performance, and create human-like interactions.

OpenAI transforms Safety Committee into independent oversight. New focus: governance, security, transparency, collaboration, unified safety frameworks for AI models.

Salesforce invests $23M in education, with $6M for AI skilling. Only 15% of workers report having necessary AI training, highlighting urgent need.

Salesforce launches Agentforce, autonomous AI agents for customer interaction. Atlas Reasoning Engine simulates human thought for complex tasks.

ChatGPT integrated with AVs interprets direct and indirect passenger commands. Tests showed lower discomfort rates and good performance even with novel instructions.

Co-LLM allows general AI to collaborate with expert models, improving factual accuracy. It uses a "switch variable" to defer to specialists at word level, enhancing efficiency.

Stanford HAI held workshop for ASEAN on AI governance. Topics: bias, fairness, trust. ASEAN plans new AI initiatives, including ethics guide and working group.

OpenAI's o1-mini matches larger models in STEM tasks at 80% lower cost. It scores 70% on AIME, reaches 86th percentile on Codeforces, and is 3-5x faster than GPT-4o.

OpenAI's new 'o1' models excel in science, coding, and math. They achieved 83% success on IMO qualifying exams, up from GPT-4's 13%.

OpenAI's o1 model excels in math, coding, and science, outperforming humans on GPQA Diamond. Uses internal "chain of thought" for complex reasoning.

OpenAI's o1 system card reveals safety assessments. Model uses chain-of-thought reasoning, shows improved safety benchmark performance.

Google's DataGemma uses Data Commons to reduce LLM hallucinations. RIG and RAG approaches improve factual accuracy in AI responses.

Anadol fine-tunes Llama models with 500M+ nature images. "Open source information is like gold for art making," he says, emphasising AI's artistic potential.

UL researchers created AI hardware using molecular-level computing. It processes info via atomic movements, achieving 4.1 TOPS/W efficiency in a 14-bit system.

Upeo Labs' Somo-GPT uses Llama 3 models as multi-subject teaching assistant in 500 Kenyan schools. Supports national curriculum digitisation efforts.

Nuro's autonomous driving system, built on NVIDIA tech, completed 1M miles with zero at-fault incidents. It now offers licensing to accelerate industry adoption.

NotebookLM's Audio Overview transforms docs into AI-hosted discussions, using Gemini 1.5 to analyse content and create engaging audio summaries.

NVIDIA and Oracle unveil zettascale OCI Supercluster with 100,000+ GPUs, delivering 2.4 zettaflops of AI compute. Boosts database AI capabilities.

Google.org funds $10M for "America's SBDC AI U," offering AI training for small businesses. AI Clinics at universities will provide hands-on instruction.

TELUS partners with Mila to explore AI technologies. Their GenAI customer support tool is the first to be certified in Privacy by Design (ISO 31700-1).

Cornell's sub-millimeter robots use kirigami to morph from 2D to 3D. 100 panels with 200 nanoscale hinges allow 40% size change via electrical stimulation.

Ada's AI Agent, using OpenAI's LLMs, doubles customer query resolution rates to 60-80%. Aims for 100% resolution in customer service.

Anthropic introduces Workspaces for Claude API, enabling granular resource management, access control, and spend limits for multiple deployments.

MIT's ScribblePrompt AI speeds up medical image analysis by 28%. Trained on 50,000+ simulations, it highlights structures via user scribbles across scan types.

Electro-hydraulic actuators mimic muscles, using oil-filled bags with electrodes. Voltage creates attraction, moving oil and shortening the bag, enabling paired movements.

DPAD AI algorithm isolates specific brain patterns from complex neural activity. It prioritizes learning behavior-related patterns, enhancing brain-computer interfaces.

AI chatbots reduced call center load by 60% in retail. In telecom, AI achieved 90% accuracy in network troubleshooting. AI fraud detection in banking cut time from 30 to 3-7 minutes.

RTX GPUs with Tensor Cores deliver up to 686 trillion AI ops/sec. New NPU offloads lightweight AI tasks. Systems run large language models 10-100x faster than CPUs.

Google.org invests in AI for global impact: $8M for AI in improving college graduation rates, $40M for responsible AI research, $10M for AI education in EMEA.

Additionally, OpenAI CEO Sam Altman will rejoin the Board of Directors, working closely with the new members and current board members Adam D'Angelo, Larry Summers, and Bret Taylor, as well as OpenAI's senior management.

ADNOC aims to become the most AI-enabled company in the energy industry.

Google upgrading Video Action to Demand Gen campaigns in 2025. Multi-format approach boosts conversions by 20%. Reaches 3B users across platforms.

Google.org launches AI initiatives for nonprofits: dedicated website, $30M funding programme, and AI-enhanced Ad Grants to boost social impact.

Deltia.ai uses NVIDIA's AI to boost manufacturing efficiency by 20%. Vision models on Jetson Orin optimise assembly lines and predict maintenance.

Claude Enterprise offers 500K token context, GitHub integration, and enhanced security. Early users report significant productivity gains across tasks.

Cornell researchers create biohybrid robots using fungal mycelia, enabling environmental sensing and responsive movement to light and touch.

LLMs show increasing covert racism towards AAE speakers. As models grow, overt racism decreases but covert racism rises, impacting decision-making systems.

Study finds 2/3 of participants let AI change life-or-death decisions, despite being warned of its limitations. Highlights dangers of overtrust in AI capabilities.

Llama models near 350M downloads, with 20M in the past month. Usage doubled from May to July 2024, and major firms are integrating Llama-based AI.

Contextual AI's RAG 2.0 integrates retriever and generator, improving AI performance 10x. It processes diverse data types with lower compute needs.

MIT study finds 70% of AI training datasets lack licensing info, 50% have errors. New tool aims to improve transparency and dataset selection.

Mila's Indigenous Pathfinders in AI program culminated with 11 participants presenting 3 AI prototypes addressing education, food, and healthcare challenges.

Infosys integrates Llama 3.1 into Topaz for document, video, and audio processing. It's also used in a legal assistant with RAG for cited information.

NVIDIA's Q2 revenue hit $30B, up 122% year-over-year. Data centre sales soared 154% to $26.3B, driven by surging demand for AI computing.

Anthropic releases Artifacts for Claude.ai across all user tiers, enabling visualisation and iteration of AI-generated work like code diagrams and dashboards.

NVIDIA's NIM Agent Blueprint integrates AI models for virtual drug screening, aiming to reduce development time and costs in pharmaceuticals.

MIT's AI + Education Summit explored AI integration in schools, featuring a climate hackathon and discussions on AI fluency as a crucial skill.

AI model provides feedback to novice peer counselors, offering objective definitions, improvement suggestions, and specific response recommendations.

Mila and Airudi's partnership aims to develop AI-driven HR solutions, enhancing recruitment, safety management, and workforce planning for 5,000+ companies.

University of Cologne's AI digital pathology platform analyses lung cancer tissue samples faster and more accurately, aiding diagnosis and treatment decisions.

Hydrogels, when connected to a virtual environment, can play Pong and improve performance over time, showing memory-like behaviour through ion movement.

This on-device AI model allows NPCs to provide quicker, more relevant responses in games. It runs locally on RTX PCs, enhancing player-character interactions.

The AI sensor, inspired by dung beetles, measures the Milky Way's orientation in low light, potentially improving navigation for drones and satellites.

NVIDIA's Mistral-NeMo-Minitron 8B model combines pruning and distillation to achieve high accuracy with 8B parameters, running on RTX workstations.

OpenAI's SearchGPT will integrate Condé Nast content, combining AI models with web sources to provide timely answers and links to original stories.

NVIDIA's StormCast AI model improves mesoscale weather prediction. Forecasts 100+ variables with 10% greater accuracy than NOAA systems. Trained on 3.5 years of climate data, aims to enhance disaster planning and mitigation efforts.

Calsoft uses NVIDIA tech for AI-powered toll automation in India. Vision AI reads plates, charges UPI accounts. Tackles night detection and weather challenges. Achieves 95% accuracy despite non-standard plates.

MIT's AI assistant uses theory of mind to monitor team beliefs, intervening to improve collaboration in tasks like search-and-rescue missions.

Google's HeAR AI model analyses cough sounds to detect diseases, trained on 300 million audio samples. It shows promise for early TB detection, outperforms other models, and is available to researchers for developing custom bioacoustic tools.

The Dresden facility, set to open in 2026, will produce 40,000 300mm wafers monthly using advanced FinFET processes, boosting Europe's chip industry.

Google Classroom marks 10 years with AI integration. LearnLM models pilot new features, enhancing personalised learning and teacher efficiency.

MIT's zinc-air microbattery, smaller than a sand grain, powers tiny robots. It uses air to oxidise zinc, generating 1 volt for circuits and sensors.

MIT researchers propose personhood credentials to verify human users online, obtained offline and used optionally to distinguish from AI accounts.

MIT's SigLLM framework uses large language models to detect anomalies in time-series data without extensive training, showing promise for complex systems.

MIT's NeuroTrALE software uses machine learning to process brain imaging data faster, reducing computing time by 90% for 32GB datasets.

MIT researchers found LLMs may form internal simulations of reality while improving language skills. A model achieved 92.4% accuracy on Karel puzzles without direct exposure, suggesting deeper understanding beyond mimicry.

Mila partners with Iter Innovandi to boost AI adoption in SMEs. Focuses on construction sector. Combines Mila's research expertise with Iter's customer focus. Aims to make advanced AI more accessible, driving innovation across Canada.

xAI released Grok-2, claiming it outperforms Claude 3.5 Sonnet and GPT-4-Turbo on LMSYS. It offers improved reasoning, real-time 𝕏 integration, and plans for multimodal features. Enterprise API access is coming soon.

California partner with NVIDIA for an AI education initiative. The program me will certify educators through NVIDIA's Deep Learning Institute, providing access to teaching materials and GPU-accelerated workstations. It aims to prepare students and adults for AI-driven careers across various sectors.

As the South African elections approach, Meta is implementing a cross platform strategy to preserve the integrity of the electoral process.

Perplexity donates $250,000 to Northwestern's Medill School for AI journalism research. Focus: AI-powered tools, improving news experiences. Aims to enhance journalistic practices while maintaining accuracy and transparency.

Rakuten partners with OpenAI to enhance services using AI. Leverages data from 1.8B users across 70 platforms. Improves customer service, shopping experience, and B2B consulting. Plans for voice and vision AI capabilities.

NVIDIA's AI tech transforms STEM education globally. GPU-accelerated curriculum in 5,000+ schools prepares students for AI-driven jobs. Offers RTX AI Toolkit, RAPIDS libraries, robotics support, and remote learning tools.

Mistral AI enables model customisation via La Plateforme and introduces AI Agents, allowing developers to create complex, shareable workflows.

The guide suggests fine-tuning, particularly parameter-efficient fine-tuning, as a more viable approach for smaller teams with limited resources compared to pre-training methods.

The guide identifies five scenarios where fine-tuning excels: customising tone and format, improving accuracy, addressing niche domains, reducing costs via distillation, and developing new abilities.

Meta AI's guide emphasises dataset quality for fine-tuning LLMs, suggesting small high-quality datasets often outperform larger low-quality ones. It compares full fine-tuning and PEFT techniques.

Google launched its AI Academy for American Infrastructure, selecting 15 startups to transform critical infrastructure using AI. The 12-week programme offers mentorship, AI training, and resources. Startups address issues like water management, transportation safety, and sustainable manufacturing.

Ornek develops AI robot guide dog IC4U using NVIDIA Jetson. Features sound sensors, 3D camera, detects objects, and assists shopping. Aims to help visually impaired navigate.

Stanford study reveals Western bias in AI chatbot alignment, compromising global effectiveness. Examined nine languages and regional dialects, finding cultural nuances lead to misunderstandings. Researchers explore causes and solutions for more inclusive AI.

Meta AI's $2 million grant programme supports innovative Llama 3.1 applications for global challenges, offering up to $500,000 per project in economic development, science, and public service.

Zoox's robotaxis, powered by NVIDIA GPUs, process data from multiple sensors. Over 5 million autonomous vehicles are projected globally by 2030.

Google expanded its AI wildfire tracking system to 15 new countries in Europe and Africa. The tool analyses satellite imagery to provide accurate fire boundaries on Google Search, Maps, and via notifications. It reached 1.4 million people in its first week of expansion.

Perplexity launches Publishers' Programme with revenue sharing, API access, and enterprise tools. Partners include TIME and Der Spiegel. Aims to align AI technology with quality journalism and expand content access.

Google's Project Green Light uses AI to optimise traffic light timing in cities worldwide. Using Google Maps data, it aims to reduce stop-and-go emissions by up to 10% at intersections. The project is live in over 70 intersections across a dozen cities.

Meta's AI Studio launches, enabling AI character creation. Zuckerberg predicts every business will have an AI. Open-source Llama 3.1 released with 405B parameters, trained on NVIDIA GPUs.

Stanford study finds cultural backgrounds shape AI preferences. European Americans prefer control, while Chinese seek connection with AI. African Americans fall between.

AI assistants for every job, says Huang. Accelerated computing saves energy. AI to transform industries from robotics to self-driving cars.

Alan Turing Institute's 2024 lecture series explores AI's impact on democracy, addressing deepfakes, bias, and trust in three expert-led discussions.

OpenAI's SearchGPT prototype combines AI models with web info to provide quick answers and clear source attribution, currently in limited testing.

The AI agent will merge Salesforce's CRM data with Workday's HR info, enabling natural language communication for easier employee task completion.

OpenAI's new Rule-Based Rewards method improves AI safety without extensive human data collection. It uses simple rules to evaluate outputs, balancing helpfulness and safety in AI models.

Mistral Large 2, a new AI model, handles 128,000 words of context and many languages. It's better at coding, math, and reasoning, with less incorrect information.

Meta releases Llama 3.1: 128K context, 8 languages, 405B model. New safety tools: Llama Guard 3, Prompt Guard, CyberSecEval 3. Open-source approach.

Meta's Llama 3.1: 405B parameter open-source AI model. 128K token context, 8 languages. New safety tools: Llama Guard 3, Prompt Guard, CyberSecEval 3.

NVIDIA launches four NeMo Retriever NIM microservices to improve LLM accuracy and performance. These tools enable efficient retrieval-augmented generation, connecting models to business data. They offer 30% fewer inaccurate answers in enterprise applications and integrate with various platforms.

NVIDIA's new NeMo Retriever microservices improve AI accuracy by efficiently fetching relevant business data. They reduce inaccurate answers by 30% in enterprise settings.

NVIDIA's AI and GPU advancements yield significant energy efficiency gains across industries, with users reporting substantial savings in finance and manufacturing.

AI expert Misha Leybovich discusses how artificial intelligence is reshaping industries and solving old problems in innovative ways.