The Gilbane Advisor

Curated for content, computing, and digital experience professionals

Month: September 2024 (Page 1 of 3)

Ai2 launches family of open multimodal models

September 25, 2024 /

The Allen Institute for AI (Ai2) announced the launch of Molmo, a family of state-of-the-art multimodal models. This family includes our best Molmo model, closing the gap between close and open models, an open, efficient, and powerful multimodal model. Currently, most advanced multimodal models can perceive the world and communicate with us, Molmo goes beyond that to enable one to act in their worlds, unlocking a new generation of capabilities, from sophisticated web agents to robotics:

  • Exceptional image understanding: Molmo can accurately understand a wide range of visual data, from everyday objects and signs to complex charts, messy whiteboards, clocks, and menus.
  • Actionable insights: To bridge the gap between perception and action, Molmo models can point to what they perceive, empowering capabilities that require spatial knowledge. Molmo can point to UI elements on the screen, enabling developers to build web agents or robots that can navigate complex interactions on screen and within the real-world.

Molmo was designed and built in the open and Ai2 will be releasing all model weights, captioning and fine-tuning data, and source code. Select model weights, inference code, and demo are available, providing open access to enable continued research and innovation in the AI community.

https://molmo.allenai.org/blog

Gilbane Advisor 9-25-24 — CORE-Bench, dynamic knowledge graphs

September 25, 2024 /

This week we feature articles from Amanda Kau, and Sayash Kapoor & Arvind Narayanan.

Additional reading comes from Ben Dickson, Fulala Fulalasen, Cassie Kozyrkov, and Paul Sawers.

News comes from Brightspot, Adobe, Hubspot, and DeepL.

Our next issue will arrive October  2.

All previous issues are available at https://gilbane.com/gilbane-advisor-index

Opinion / Analysis

Can AI automate computational reproducibility?

Sayash Kapoor and Arvind Narayanan tackle a critical issue…
“In a new paper, we introduce CORE-Bench (Computational Reproducibility Agent Benchmark), a benchmark for measuring how well AI can automate computational reproducibility, that is, reproducing a paper’s findings when the code and data are available.”

Introduction: (5 min) 
https://www.aisnakeoil.com/p/can-ai-automate-computational-reproducibility
Full paper:
PDF: https://arxiv.org/pdf/2409.11363v1
HTML: https://arxiv.org/html/2409.11363v1

Dynamic knowledge graphs: a next step for data representation?

How do you integrate temporal data into knowledge graphs? What’s the difference between dynamic knowledge graphs, temporal knowledge graphs, and discrete time dynamic graphs? What kinds of applications are they used for? And how do they work, or compete with, LLMs? Amanda Kau does a nice job addressing these and other questions. (9 min)

https://medium.com/@researchgraph/dynamic-knowledge-graphs-a-next-step-for-data-representation-c35a205a520a

More Reading

All Gilbane Advisor issues

Content technology news

Brightspot announces latest version of their content management system

Updated CMS has expansive tech stack integrations, AI features that speed up creation time, and prioritizes streamlined content workflows.
https://www.brightspot.com

HubSpot launches Breeze, updates Marketing & Content Hubs

Breeze includes Copilot, an AI companion; Agents, AI experts who complete work from start to finish; 80+ AI features embedded across the platform.
https://www.hubspot.com/company-news/fall24-spotlight

DeepL announces glossary generator

Companies can create and manage custom translation glossaries so specific words or phrases are translated consistently according to their unique terminology.
https://www.deepl.com/en/translator

Adobe Experience Cloud updates help personalize and measure AI-generated content

New offerings enable brands to personalize AI-generated content variations across channels such as web and email to boost customer engagement.
https://news.adobe.com/news/news-details/2024/Adobe-Unveils-Innovations-in-Adobe-Experience-Cloud-for-Brands-to-Personalize-and-Measure-AI-generated-Content/default.aspx

All content technology news

The Gilbane Advisor is authored by Frank Gilbane and is ad-free, cost-free, and curated for content, computing, web, data, and digital experience technology and information professionals. We publish recommended articles and content technology news most Wednesdays. We do not sell or share personal data.

Subscribe | View online | Editorial policy | Privacy policy | Contact

Contentstack adds visual building capabilities to CMS

September 24, 2024 /

Contentstack, a composable digital experience platform (DXP) and Enterprise Headless CMS, announced native visual building capabilities to its platform. Visual Builder’s experience-building capabilities accelerate workflows for marketers. Marketers can now see their work, present and future, in a headless environment. With Contentstack, teams can now natively visualize:

  • Composability across vendors (e.g. pulling content from a DAM + CMS)
  • Personalization and audience impersonation
  • Generative AI authoring
  • Future iterations of their digital properties

As more organizations adopt Headless CMS technology and composable architectures to support the demands of modern marketing, the need for business users to be able to visualize their work has steadily increased. Digital experience creation moved to the back end, leading to complicated planning and trial-and-error processes. Contentstack’s Visual Builder offers a full set of tools for users to create and see their work in real time, across any persona.

Developers and designers will also be able to build components and layouts with drag-and-drop functionality, maintaining firm guardrails around how a content author’s work will look.

https://www.contentstack.com/platforms/visual-building

Salesforce to acquire Zoomin

September 24, 2024 /

Salesforce has signed an agreement to acquire Zoomin, a data management provider for unstructured data. Salesforce Data Cloud, with Zoomin’s capabilities, will accelerate the unlocking of unstructured data to power AI agents, making them more personalized and context-aware in every customer interaction. With this, Salesforce’s Agentforce will gain new levels of intelligence, enabling customers to build AI that provides real-time, data-informed responses and actions tailored to individual customer needs. The acquisition is expected to close in the fourth quarter of Salesforce’s fiscal year 2025, subject to customary closing conditions.

With Zoomin augmenting Data Cloud, Salesforce can power new use cases for customers across multiple customer touch points. As an example, Service Cloud customers will be able to leverage enterprise knowledge to automate service interactions and improve Agentforce assistance for service representatives. By leveraging enterprise knowledge, service teams will enable humans to work with agents to achieve faster resolutions, higher customer satisfaction, and lower operational costs. Knowledge is crucial for customer success, and because knowledge is distributed, Zoomin enables the integration of unstructured data from various sources, empowering service teams with a comprehensive understanding of their customers.

Salesforce customers are already using Zoomin’s unified knowledge product to deliver intuitive experiences.

https://www.salesforce.com/news/stories/salesforce-signs-definitive-agreement-to-acquire-zoomin/

MindsDB launches conversational enterprise-ready AI

September 24, 2024 /

MindsDB announced the launch of Minds, a turnkey, private conversational AI designed to reason and orchestrate across real-time and historical enterprise data to answer questions for AI applications and agents.

With Minds, developers can move from prototyping to building production-ready AI applications in weeks, without requiring specialized AI expertise. These applications can reason and understand user requests in natural language, including use cases for conversational search, analytics, and triage.

Minds answer questions across complex and diverse data environments. Minds perform transparent semantic query planning, showing users how they “think” through answering questions, as follows:

  • Minds are natural language fluent and understand user intent through conversation.
  • Based on that intent, Minds select the best data sources across databases, data warehouses, file systems, and applications, providing a unified view of an organization’s data.
  • Minds then translate the question into a plan of queries, optimized and federated across the chosen data sources.
  • Minds are also capable of applying calculations, such as sorting or structuring, needed to deliver the right answer.

Available as a turnkey, private solution, Minds ensure security, privacy, and regulatory compliance. Minds operate securely across any cloud or on-premises environment via Nvidia NIMs, and can be deployed serverlessly in MindsDB’s cloud.

https://mindsdb.com/newsroom/mindsdb-launches-conversational-enterprise-ready-ai-that-shows-you-how-it-thinks

HubSpot launches Breeze, updates Marketing & Content Hubs

September 18, 2024 /

HubSpot launched Breeze, its embedded AI with new Copilot and Agents; Breeze Intelligence for data enrichment; and updates to Marketing Hub and Content Hub. Featured releases:

  • Breeze, HubSpot’s AI to power the customer platform. Breeze includes:
    • Copilot, HubSpot’s AI companion to make work easier.
    • Four Breeze Agents to get work done fast, including Content Agent, Social Media Agent, Prospecting Agent and Customer Agent.
  • Breeze Intelligence, HubSpot’s data enrichment and buyer intent solution:
    • Data enrichment pulls from a database of over 200 million buyer and company profiles to enrich company and contact records in HubSpot’s Smart CRM.
    • Buyer intent to help customers identify which prospects are best fit.
    • Form shortening to increase conversion by automatically adding information Breeze Intelligence already knows.
  • Updates to Marketing Hub and Content Hub including tools:
    • to capture attention like Content Remix for video, which uses AI to turn a single video into a full campaign of clips, audio and written content.
    • like Lead Scoring to find high engagement, high-fit prospects, and Google Enhanced Conversions to leverage first-party conversion data from HubSpot.
    • to measure impact like the new Marketing Analytics Suite, which brings all marketing metrics and reporting in one place and get results faster.

https://www.hubspot.com/company-news/fall24-spotlight

Gilbane Advisor 9-18-24 — o1 paradigm?, abductive reasoning

September 18, 2024 /

This week we feature articles from Alberto Romero, and Tony Seale

Additional reading comes from Mike Knoop, Emma Roth, Frederic Lardinois, and Michael Nuñez.

News comes from Cloudera, Adobe, Brightcove, and SearchStax & Magnolia.

Our next issue will arrive September 25.

All previous issues are available at https://gilbane.com/gilbane-advisor-index

Opinion / Analysis

OpenAI o1: A new paradigm for AI

Alberto Romero provides a high-level yet in-depth explanation of OpenAI’s latest model that “reasons” and opines on what it means near and long-term. In spite of the presence of “paradigm” in the title, Romero does a nice job balancing enthusiasm and skepticism. A thoughtful read. (19 min)

https://www.thealgorithmicbridge.com/p/openai-o1-a-new-paradigm-for-ai

LLMs, faking, it and abductive reasoning

Tony Seale’s short post on three types of reasoning is good companion piece to Romero’s…

“A practical implementation of this [Ed.: abductive reasoning] might integrate the inductive curve-fitting capabilities of large language models with the formal inference mechanisms of knowledge graphs, creating a dynamic neural-symbolic loop.” (2 min – 9 including comments) 

https://www.linkedin.com/posts/tonyseale_the-question-of-whether-large-language-models-activity-7237731298952302594-T0mN

More Reading

All Gilbane Advisor issues

Content technology news

Cloudera adds Accelerators for Machine Learning Projects (AMPs)

AMPs are end-to-end machine learning (ML) based projects that can be deployed with a single-click directly from the Cloudera platform.
https://www.cloudera.com/about/news-and-blogs/press-releases/2024-09-12-cloudera-unveils-new-suite-of-accelerators-for-machine-learning-projects-amps.html

Adobe announces Adobe Express updates and special teams offer

New capabilities help SMBs create on-brand content across Marketing, HR and Sales accelerate content creation and work across teams.
https://news.adobe.com/news/news-details/2024/Adobe-Express-Updates-Deliver-More-Value-for-Solopreneurs-and-SMBs-with-Innovation-and-Special-Teams-Offer/default.aspx

Brightcove unveils Brightcove AI Suite

The Suite includes: AI Content Multiplier, AI Universal Translator, AI Metadata Optimizer, AI Engagement Maximizer, and AI Cost-to-Quality Optimizer.
https://campaigns.brightcove.com/ai-solutions/

SearchStax and Magnolia partner on personalized search solutions

Partnership helps marketers and developers by merging the search capabilities of SearchStax Site Search with Magnolia’s enterprise-grade DXP.
https://www.searchstax.com ■ https://www.magnolia-cms.com

All content technology news

The Gilbane Advisor is authored by Frank Gilbane and is ad-free, cost-free, and curated for content, computing, web, data, and digital experience technology and information professionals. We publish recommended articles and content technology news most Wednesdays. We do not sell or share personal data.

Subscribe | View online | Editorial policy | Privacy policy | Contact

DeepL announces glossary generator

September 18, 2024 /

DeepL, a global Language AI company, announced updates to its glossary feature, which is a tool that enhances translation consistency and accuracy by enabling professionals and companies to personalize translations for specific terms. The tool now offers a smart glossary generator, which helps simplify and speed up the process of creating glossaries for translations. Expanded capabilities:

  • An AI-powered glossary generator: DeepL’s new glossary generator is a tool enabling teams to create custom translation glossaries with a simple file upload. Previously translated files can be leveraged to generate entries for personalized DeepL glossaries, reducing manual work and enhancing efficiency.
  • Expanded glossary language functionality: Glossary now supports Korean, Danish, Swedish, Norwegian, and Romanian translations, bringing the total number of languages to 16.
  • More convenient access across DeepL platforms: Users can access and apply the glossary within the DeepL browser extensions for Chrome and Edge, enabling consistent translations across the web, including Google Workspace applications. The glossary can also be applied within DeepL for Microsoft 365 integrations, including Word, Outlook, and PowerPoint, and be accessed through DeepL’s web browser, desktop apps, and API.

DeepL’s glossary generator is now available to Pro Advanced and Ultimate subscribers with support for TMX/DOCX/PDF files.

https://www.deepl.com/en/translator

« Older posts

Policies

Contact

twitterrssMastadonBluesky

Contact • customerservice@gilbane.com

© 2024 Bluebill Advisors (except for contributed posts which are licensed under creative commons)

Banner photos: Fort Point Channel, Summer St bridge, and Northern Ave bridge, Boston

Bluebill Advisors Inc logo

© 2024 The Gilbane Advisor

Theme by Anders Noren — Up ↑