Datagrom AI News Logo
Sep 25, 2024
Datagrom AI News Cloud

GPT-generated summaries

Archived stories
Older than 7 days
Updated daily
118 archived articles (Page 5 of 6)

OpenAI reportedly preparing to launch new ‘Strawberry’ AI model

OpenAI reportedly preparing to launch new ‘Strawberry’ AI model

August 27, 2024: OpenAI Set to Release Advanced Strawberry AI Model - OpenAI is preparing to launch a new AI model named Strawberry, previously known as Q*, which can solve advanced math problems, perform high-level tasks, and tackle complex puzzles. Scoring over 90% on the MATH benchmark, Strawberry surpasses current models like GPT-4 and GPT-4o. Previously, concerns about its capabilities briefly led to the ouster of CEO Sam Altman, highlighting its potential risks. Strawberry may mark a significant step toward artificial general intelligence and is expected to release in the fall.

Five months after Microsoft hired its founders, Inflection adds usage caps to Pi

Five months after Microsoft hired its founders, Inflection adds usage caps to Pi

August 26, 2024: Inflection Caps Pi Usage, Eyes Enterprise Market - Inflection will cap free access to its AI chatbot Pi while allowing users to export conversations. This shift comes after Microsoft hired its founders and licensed Inflections AI model for $650 million, triggering antitrust probes. CEO Sean White is steering the company toward enterprise products, citing resource limitations. Inflection aims to set industry standards in data mobility and transferability, with significant interest from organizations for API access. White is in talks with multiple large companies and plans to launch enterprise products this fall.

Anthropic publishes the ‘system prompt’ that makes Claude tick

Anthropic publishes the ‘system prompt’ that makes Claude tick

August 26, 2024: Anthropic Unveils Claudes System Prompts for Transparency - Anthropic has publicized the system prompts guiding its latest Claude AI models, highlighting transparency and ethical AI. These prompts define behavior and traits, ensuring models like Claude 3.5 Opus remain polite, intellectually curious, and impartial. They also strictly restrict certain actions, such as facial recognition. This move aims to contrast Anthropic from other firms that keep such details secret, and represents an effort to align AI interactions closely with ethical standards.

OpenAI, Adobe and Microsoft support California bill requiring watermarks on AI content

OpenAI, Adobe and Microsoft support California bill requiring watermarks on AI content

August 26, 2024: Tech Giants Back California AI Watermark Bill - OpenAI, Adobe, and Microsoft support Californias AB 3211, requiring watermarks on AI-generated content, now heading for a final vote. The bill mandates clear labeling on platforms such as Instagram and X. Amendments to the bill swayed previous opponents, including a trade group of large software makers.

Viggle AI raises $19M in funding for its AI video generator

Viggle AI raises $19M in funding for its AI video generator

August 26, 2024: Viggle AI Secures $19M for AI Video Tech - Viggle AI, known for its generative AI video service, has raised $19 million in Series A funding led by Andreessen Horowitz. The funds will be used to enhance product development and market reach. Viggle AIs unique JST-1 model is lauded for its realistic character movements, setting it apart from competitors like OpenAI and Runway AI. With over four million users, the company plans to grow its team and offer new features via a creator program.

Anthropic publishes the ‘system prompts’ that make Claude tick

Anthropic publishes the ‘system prompts’ that make Claude tick

August 26, 2024: Anthropic Releases Claudes System Prompts - Anthropic has made public the system prompts for its AI models, Claude 3.5 Opus, Sonnet, and Haiku, showcasing its commitment to transparency. These prompts guide the AIs behavior, outline banned capabilities like facial recognition, and define desired personality traits such as intellectual curiosity and impartiality. This unprecedented disclosure could pressure competitors to follow suit.

Meta and Spotify CEOs criticize AI regulation in the EU

Meta and Spotify CEOs criticize AI regulation in the EU

August 23, 2024: Meta and Spotify CEOs Slam EU AI Rules - Meta CEO Mark Zuckerberg and Spotify CEO Daniel Ek criticized EU privacy regulations for stifling AI innovation, claiming it blocks European access to advanced AI models and hinders cultural representation. Meta announced it will withhold its new multi-modal AI model, Llama, from the EU due to regulatory uncertainties. Spotify highlighted how open-source AI could enhance artist discovery and streaming, advocating for simplified regulations to support European tech growth. The companies previously supported EU regulations against Apples App Store practices but now argue current AI laws are overly restrictive.

‘Disappointed but not surprised’: Former employees speak on OpenAI’s opposition to SB 1047

‘Disappointed but not surprised’: Former employees speak on OpenAI’s opposition to SB 1047

August 23, 2024: Former OpenAI Researchers Criticize AI Bill Opposition - Ex-OpenAI researchers Daniel Kokotajlo and William Saunders denounce OpenAIs stance against Californias AI disaster prevention bill SB 1047, calling it contradictory to CEO Sam Altmans previous support for AI regulation. They hope regulation will steer OpenAI towards safer AGI development. Rival Anthropic supports the bill, citing its overall benefits after amendments.

Beyond OpenAI: The rise of not-too-large language models

Beyond OpenAI: The rise of not-too-large language models

August 23, 2024: Summary - This week saw a trend toward smaller vertical-specific AI models, with notable releases from Nvidia, Microsoft, and OpenAI. Chinese startups are also adopting smaller models to bypass energy concerns and export controls. Key earnings reports had mixed results; Snowflake fell, while others like Zoom and Synopsys showed gains. AI and data continue to evolve with specialized and fine-tuned models, attracting enterprise interest for cost and privacy benefits. In other news, significant acquisitions and funding rounds emphasized ongoing innovation in AI and data infrastructure. HeadlineSmaller AI Models Gain Ground as Earnings Vary

Open source Dracarys models ignite generative AI fired coding

Open source Dracarys models ignite generative AI fired coding

August 23, 2024: Dracarys Models Turbocharge AI Coding Efficiency - Abacus.ai introduces Dracarys, a new suite of open-source large language models (LLMs) optimized for coding tasks. Using a combination of specialized training datasets and fine-tuning techniques, Dracarys significantly enhances the performance of models like Metas Llama 3.1 and Qwen2. Notably outperforming existing benchmarks, Dracarys aims to benefit developers and enterprises who prefer not to rely on public APIs.

Beyond OpenAI: The rise of not-too-large language models

Beyond OpenAI: The rise of not-too-large language models

August 23, 2024: Rise of Vertical AI Models - Smaller, industry-specific AI models, dubbed VLMs, are gaining traction as companies like Nvidia, Microsoft, and OpenAI focus on efficiency and data governance. Chinese startups also explore these models to circumvent Nvidia GPU export controls. Major corporate and earnings movements include AMDs acquisition of ZT Systems, Grafana Labs $270M funding, and Snowflakes stock drop. New AI models and services from major players like Salesforce, Waymo, and Google Cloud were announced. Additionally, significant cyberattacks and regulatory updates further shaped the tech landscape this week.

Beyond OpenAI: The rise of not-too-large language models

Beyond OpenAI: The rise of not-too-large language models

August 23, 2024: ## VLMs Rise: Tailored AI Models Gain Traction - Smaller language models designed for specific industries are gaining momentum. Nvidia and Microsoft have released new VLMs, with OpenAI offering GPT-4o fine-tuning. Chinese firms adopt cloud solutions to bypass export controls. AMD acquires ZT Systems to compete with Nvidias GPUs. Infrastructure observability and other tech sectors see substantial funding and positive earnings reports. However, Snowflakes shares drop due to revenue concerns. Notably, Autonomys Mike Lynch tragically died at sea. Upcoming events include VMware Explore and key earnings announcements from numerous tech giants.

Beyond OpenAI: The rise of not-too-large language models

Beyond OpenAI: The rise of not-too-large language models

August 23, 2024: AI Evolution: Rise of Custom Language Models - The AI landscape is shifting towards smaller, application-specific language models (VLMs). Major players like Nvidia, Microsoft, and OpenAI are releasing refined, targeted models catering to enterprise needs for privacy and efficiency. China is using cloud services to bypass export controls for advanced GPUs. Significant acquisitions and funding rounds highlight the growing focus on AI infrastructure, with companies such as AMD, Grafana Labs, and Redis announcing major developments. Meanwhile, earnings were mostly positive, despite a few setbacks like Snowflakes revenue outlook. SiliconANGLE and theCUBE will cover upcoming tech events and earnings reports.

Salesforce’s newest AI agents help to filter out sales prospects and train salespeople

Salesforce’s newest AI agents help to filter out sales prospects and train salespeople

August 22, 2024: Salesforce Unveils New AI Sales Helpers - Salesforce introduces two AI agents, Einstein SDR and Einstein Sales Coach, to enhance its CRM platform. The Einstein SDR Agent autonomously engages and filters sales prospects using agentic AI and RAG techniques, minimizing the need for human intervention. The Einstein Sales Coach provides personalized coaching through lifelike role-playing scenarios, improving sales pitches and negotiation skills. Both AI tools, built on the Einstein 1 Agentforce platform, will be available to Salesforce customers in October and can be customized using no-code actions and prebuilt templates.

Opkey, an AI-based ERP testing platform, raises $47M

Opkey, an AI-based ERP testing platform, raises $47M

August 22, 2024: Headline - Opkey Raises $47M for AI ERP Testing SummaryOpkey, an AI platform for testing ERP software, raised $47M in Series B funding, led by PeakSpan Capital. The startup, co-founded by ERP veterans from Adobe and Oracle, helps over 200 enterprises ensure seamless integration and operation of ERP systems like Oracle, Workday, and Salesforce. With growing enterprise SaaS adoption and the need for continuous testing, Opkey addresses critical vulnerabilities and enhances user experience. This funding aims to extend Opkeys market reach as ERP spending continues to rise.

Harmonyze wants to build AI agents to help franchisors make sense of unstructured data

Harmonyze wants to build AI agents to help franchisors make sense of unstructured data

August 22, 2024: Harmonyze AI Simplifies Franchise Management - Brooklyn-based startup Harmonyze raised $2M in a pre-seed round to develop AI agents that help franchisors manage unstructured data for compliance and operational efficiency. Founded by Gary Liskovich and Jonny Greenspan in 2023, Harmonyze targets franchisors with extensive networks. The AI agents can perform over 200 tasks, reducing time on administrative duties. The company plans to use the funds for hiring and product development, focusing on feedback from a select group of initial customers.

AI21 Labs’ updated hybrid SSM-Transformer model Jamba gets longest context window yet

AI21 Labs’ updated hybrid SSM-Transformer model Jamba gets longest context window yet

August 22, 2024: AI21 Labs Unveils Enhanced Jamba LLMs with Longest Context - AI21 Labs introduced Jamba 1.5 Mini and Large, open-source hybrid SSM-Transformer models with 256,000 token context windows. These models efficiently handle longer data sequences, providing superior performance and latency over competitors like Llama 8B and 70B by leveraging a unique architecture that marries Transformers with Mambas structured state space. This results in improved context understanding, essential for enterprise AI tasks and reducing costs.

Agribusiness AI startup Ceres Imaging rebrands as Ceres AI after closing on late-stage funding

Agribusiness AI startup Ceres Imaging rebrands as Ceres AI after closing on late-stage funding

August 22, 2024: Ceres Imaging Rebrands to Ceres AI - Agribusiness AI startup Ceres Imaging has rebranded as Ceres AI after a Series D funding round led by Remus Capital. The company specializes in precision farming through high-resolution aerial imagery and AI-driven data analytics, helping farmers improve crop yields and sustainability. Now expanding into financial services, Ceres AI aims to license its vast agribusiness data to insurers and lenders. The funding will support growth into vertical markets within agribusiness and financial sectors, with Remus Capital CEO Krishna K. Gupta appointed as the new chairman.

Opkey, an AI-based ERP testing platform, raises $47M

Opkey, an AI-based ERP testing platform, raises $47M

August 22, 2024: Opkey Raises $47M for AI-Powered ERP Testing - Opkey, an AI-driven platform for continuous ERP software testing, secured $47 million in a Series B round led by PeakSpan Capital. With over 200 large enterprises as clients and partnerships with major integrators like KPMG and PwC, Opkey aids organizations by automating the testing of ERP systems such as Oracle, Workday, and Salesforce. Co-founded by ERP industry veterans, Opkey seeks to mitigate integration issues exacerbated by continuous cloud updates, addressing a critical market gap and competing with players like Leapwork and Katalon.

Google Cloud Run speeds up on-demand AI inference with Nvidia’s L4 GPUs

Google Cloud Run speeds up on-demand AI inference with Nvidia’s L4 GPUs

August 21, 2024: Google Cloud Run Integrates Nvidia L4 GPUs - Google Cloud Run now supports Nvidia L4 GPUs, enabling rapid on-demand AI inference for generative models and other AI workloads. Available in preview in select regions, it offers fast token rates for models with up to 9 billion parameters, automatic scaling, and pay-per-use pricing. This enhancement reduces latency and ensures efficient resource use, benefiting applications such as chatbots, image generators, and video streaming. Early adopters like LOral report impressive low latency and responsiveness.