Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More When OpenAI released desktop app versions of ChatGPT, it was clear the goal was to get more users to bring ChatGPT into their daily workflows. Now, new updates to Mac OS and Windows PC versions encourage users to stay in the ChatGPT apps for most of their tasks.  Some ChatGPT on Mac OS users can now open third-party applications directly from the app. ChatGPT Plus and Teams subscribers — with ChatGPT Enterprise and Edu users following soon after — can access VS Code, Xcode, Terminal and iTerm2 from a dropdown.  This kind of integration calls to mind GitHub Copilot’s integration with coding platforms announced in October.  Alexander Embiricos, product lead with the ChatGPT desktop team, said one of the biggest user behaviors the company saw was copy-pasting text or code generation with ChatGPT to other applications. Embiricos was the CEO of Multi, a screen sharing and collaboration startup acquired by OpenAI in June. “We wanted to start integration with [integrated development environments] IDEs because we know a lot of our customers are developers, as we were seeing a lot of copy-pasting text-based material from the app to other platforms,” Embiricos said.  He added that OpenAI wanted to focus on privacy while building the integrations, so the third-party apps would only open manually.  Users can begin coding with ChatGPT and choose VS Code from the app. Once launched, VS Code will open with the same code that they were working on. Embiricos said theoretically, people can have multiple third-party apps open while using ChatGPT on Mac.  Right now, third-party app integration is only available on Mac OS, but Embiricos said PC users will also get the feature eventually. OpenAI also plans to expand the number of apps in the future.  Windows PC is not left behind The Windows PC version of the ChatGPT desktop app will now be available for download to all ChatGPT users, following the limited release to subscribers. Along with expanding the user base, OpenAI updated the PC app with access to Advanced Voice Mode and screenshot capabilities.  Embiricos said customers have asked them to use Advanced Voice Mode on desktop for a while, so they wanted to focus on the feature for the PC app. The screenshot capability will also take advantage of some specific features in Windows machines, which will let users choose which windows to take a photo of.  “ChatGPT can understand what you’re describing to it, of course, but if you add a photo to your chat, its responses are richer, and we see a lot of users copy-pasting photos into ChatGPT so adding a screenshot option makes that easier,” Embiricos said. Many of the features in the Mac OS desktop app will also come to PC, but Embiricos noted that the team focused on making the PC app more widely available first.  Interfaces are the new battle ground Chat interfaces like ChatGPT proved incredibly useful to a variety of users, but before the advent of desktop versions, people had to go to a website to generate text or code or photos and have to bring chat responses to whichever application they’re doing actual work with.  So it’s no surprise that companies like OpenAI want to capture more of their customer base by bringing their workflows closer to their interface. GitHub made this possible with its integrations with VS Code and Xcode. Anthropic’s Claude, while not integrated with third-party apps, created Artifacts so users don’t have to go elsewhere to see what their generated webpage looks like. OpenAI followed suit with Canvas, which functions similarly.  Meanwhile, Amazon Web Services (AWS) just made its Q Developer AI assistant integrated into popular IDEs Visual Studio Code and JetBrains as an in-line suggestions and code completion add-on, allowing them to highlight chunks of their code and type instructions directly into the LLM without toggling over to another screen. App integration is nothing new for software, as many companies often work together to bring services to where users are. For example, Slack includes apps from Zoom, Atlassian, Asana, and Google that people can call up within a chat window.  source

The hype around generative AI may be starting to cool, according to a new Slack report. The survey of more than 17,000 desk workers worldwide, published on Nov. 12, revealed a disconnect between AI aspirations and adoption rates. The report focused on barriers to AI adoption at work and how leaders can clarify questions about it. The slight dip in global interest is notable after nearly a year of increased excitement around AI. “AI adoption isn’t just about enterprises, it’s also about employees,” said Christina Janzer, head of Slack’s Workforce Lab, in a press release. “With sentiment around AI dropping, businesses need to help employees accelerate their AI journey and address the cultural and organizational blockers standing in their way.” AI may not be fulfilling the hype Slack saw a steady rise in AI adoption from September 2023 to March 2024. At its peak in March 2024, about one-third (32%) of desk workers surveyed had used AI to do their jobs. This percentage began to show a decline or plateau in the last three months. Specifically: The U.S. saw just a single percentage point of growth in AI usage, from 32% to 33% of desk workers. “Excitement” about AI helping with work tasks among global workers dropped by 6%. Excitement about AI fell 9% over the last three months in the U.S. Excitement about AI fell 12% over the last three months in France. Nearly all (99%) of surveyed executives say they will make an investment into AI this year. More must-read AI coverage Nearly half of employees would be uncomfortable revealing AI use to managers While some companies create top-down initiatives to encourage the use of AI, many employees are reluctant to share their AI use: 48% of survey participants said they would be uncomfortable telling their managers they use AI. They feared a perception of AI use as cheating, a resource of the less competent, or laziness. Notably, Slack asked participants whether they would be uncomfortable sharing their AI use with their manager, not whether they are uncomfortable using AI at all. Survey-takers who are comfortable sharing that they use AI at work are likelier to use it. Still, the underlying fears reflect on both the technology and company culture. SEE: AI can introduce security risks to organizations and security teams. In general, the usage of much-hyped technologies tends to level out over time. Slack noted that the potential “lazy” and “cheating” accusations, the perception that AI is “not yet living up to the hype,” and a lack of training in using AI are the primary factors that affect employees’ viewpoints of the technology. Employees are concerned AI will not reduce administrative tasks AI advocates have long argued that the technology helps companies by automating rote tasks, thereby freeing up time for meaningful activities that support an organization’s bottom line. However, Slack’s report indicated that organizations haven’t seen reduced administrative tasks over the past few months. Instead, many employees suspect AI could lead to more drudge work and an increased workload. “Employees are worried that the time they save with AI will actually increase their workload — with leaders expecting them to do more work, at a faster pace,” Janzer said. “This presents an opportunity for leaders to redefine what they mean by ‘productivity,’ inspiring employees to improve the quality of their work, not just the quantity.” When asked what they would want to do with the time saved by AI, participants said they wanted time to engage in non-work activities and skill-building. But when asked what they would likely do with extra time, people listed administrative tasks and additional work on existing projects. How team leaders who want to promote AI can change minds Slack — which has its own AI assistant — recommends managers who want to promote AI: Run team-building exercises related to AI. Make AI use, and AI wins, visible to the entire organization through convenient communications channels. Model AI use in managers’ own work as appropriate. Focus on skill-building and training in how to use generative AI. Redefine what productivity means, tying goals to innovative or creative work projects to incentivize giving the drudge work to the AI. Remember that AI can’t replace real human connection. Approach your team’s connections and the ways people ask one another for assistance with “intentionality.” Slack also recommended that organizations train their employees to use AI via short, impactful sessions — also known as “microlearning.” “AI training programs don’t have to be a heavy lift,” said Chrissie Arnold, director of future of work programs at Workforce Lab. “At Slack, we’ve had pretty amazing results from just 10 minutes a day of AI microlearning.” source

The data and AI industries are constantly evolving, and it’s been several years full of innovation. Even less experienced technical professionals can now access pre-built technologies that accelerate the time from ideation to production. As a result, employers no longer have to invest large sums to develop their own foundational models. They can instead leverage the expertise of others across the globe in pursuit of their own goals. However, the road to AI victory can be bumpy. Such a large-scale reliance on third-party AI solutions creates risk for modern enterprises. It’s hard for any one person or a small team to thoroughly evaluate every tool or model. Yet, today’s data scientists and AI engineers are expected to move quickly and create value. The problem is that it’s not always clear how to strike a balance between speed and caution when it comes to adopting cutting-edge AI. As a result, many companies are now more exposed to security vulnerabilities, legal risks, and potential downstream costs. Explainability is also still a serious issue in AI, and companies are overwhelmed by the volume and variety of data they must manage. Data scientists and AI engineers have so many variables to consider across the machine learning (ML) lifecycle to prevent models from degrading over time. It takes a highly sophisticated ML operation to build and maintain effective AI applications internally. The alternative is to take advantage of more end-to-end, purpose-built ML solutions from trusted enterprise AI brands. Introducing Cloudera AMPs To help data scientists and AI engineers, Cloudera has released several new Accelerators for LL Projects (AMPs). Cloudera’s AMPs are pre-built ML prototypes that users can deploy with a single click within Cloudera The new AMPs address common pain points across the ML lifecycle and enable data scientists and AI engineers to launch production-ready ML use cases quickly that follow industry best practices. Rather than pursue enterprise AI initiatives with a combination of black box ML tools, Cloudera AMPs enable companies to centralize ML operations around a trusted AI leader. They reduce development time, increase cost-effectiveness for AI projects, and accelerate time to value without incurring the risks typically associated with third-party AI solutions. Each Cloudera AMP is a self-contained prototype that users can deploy within their own environments and are open-source projects, demonstrating the company’s commitment to serving the broader open-source ML community. Let’s dive into Cloudera’s latest AMPs: The PromptBrew AMP is an AI assistant designed to help AI engineers create better prompts for LLMs. Many developers struggle to communicate effectively with their underlying LLMs, so the PromptBrew AMP bridges this skill gap by giving users suggestions on how to write and optimize prompts for their company’s use cases.  RAG with Knowledge Graph on CML The RAG with Knowledge Graph AMP showcases how using knowledge graphs in conjunction with Retrieval-augmented generation can enhance LLM outputs even further. RAG is an increasingly popular approach for improving LLM inferences, and the RAG with Knowledge Graph AMP takes this further by empowering users to maximize RAG system performance.  Chat with Your Documents The Chat with Your Documents AMP allows AI engineers to feed internal documents to instruction-following LLMs that can then surface relevant information to users through a chat-like interface. It guides users through training and deploying an informed chatbot, which can often take a lot of time and effort. Lastly, the Fine-tuning Studio AMP simplifies the process of developing specialized LLMs for certain use cases. It allows data scientists to focus pre-existing models around specific tasks within a single ecosystem to manage, refine, and evaluate LLM performance. A clearer path to ML success With Cloudera AMPs, data scientists and AI engineers don’t have to take a leap of faith when adopting new ML tools and models. They can lean on AMPs to mitigate MLOps risks and guide them to long-term AI success. AMPs are catalysts to fast-track AI projects from concept to reality with pre-built solutions and working examples, ensuring that use cases are dependable and cost effective while reducing development time. Businesses no longer need to pour time and money into building everything in-house, companies can move fast in today’s hyper-competitive business landscape. For more on Cloudera’s AMPs, click here. source

Data ingestion into security information and event management (SIEM) is too expensive. In fact, it’s so expensive that “How do we reduce our SIEM ingest costs?” is one of the top inquiry questions I get from Forrester clients. And the problem is not new — security leaders have struggled with managing their SIEM budget for over a decade. Visibility without actionability is an expensive waste of time. The increasing spend in SIEM is driven by a few factors. First, the shift to the cloud produced more data to intake and store. To scale at the rate of ingest, SIEM vendors moved their offerings to the cloud — a shift that necessitated ingest-based pricing to balance out cost. But most importantly, the crux of SIEM cost challenges stems from the belief that more data in the SIEM is better. Security is a big data problem, right? More data, more visibility, more insights … right? Not quite. Data — and subsequent visibility into that data — is meaningless without actionability. Data is brought into the SIEM for compliance requirements and for alerting on potential attacker activity. To alert on attacker activity, a human being needs to build a rule. Visibility into the data is only half the battle. You could have all the visibility in the world, but without those rules, you will not find the attackers consistently and in a more automated way. Instead, we recommend focusing what you ingest on what’s most important for compliance and alerting. But it isn’t always easy to do so because: Logs have extra fields that you don’t always need. The structure changes and is different between vendors. You want some logs to go to a certain datastore with others elsewhere. You may want to redact data for privacy reasons. Further, indexed data can often become 3–5x the original size. SIEM vendors have the ability to address some of these challenges, but the capabilities tend to be limited and cumbersome to use. The vendors haven’t created effective tools for log size reduction or routing especially, since it directly opposes their own interests: getting you to ingest more data into their platform and, therefore, spend more money with them. Data pipeline management tools reduce data preparation. This is where data pipeline management (DPM) tools for security come in. DPM tools can route, reduce, redact, enrich, or transform data. The benefits of a purpose-built data pipeline tool are to reduce the data preparation necessary to interpret the streams of data and events specific to security insights. With increasingly distributed and disparate systems, a purpose-built data pipeline tool is designed to address complexity of classification, integration, and modeling data for analysis. Security teams get immediate value from its ability to reduce log sizes and thus ingest costs. In the longer term, however, much of the value comes from storage tiering or data routing — being able to redirect data to the storage location of your choice. For example, short-term data valuable for incident response can be routed directly to extended detection and response (XDR), while data for compliance requirements can be directed to longer-term, cheaper storage. This can be useful across the business, especially for those that have data storage requirements for different use cases such as compliance, detection and response, or IT. When it comes to DPM tools for security, Cribl is one of the earliest to market and the most ubiquitous, but others such as Tenzir, Tarsal, DataBahn, Calyptia, observIQ, and Observo AI are also built to manage data pipelines for security use cases. Some SIEM and XDR vendors are also building more robust data pipeline management capabilities, like Splunk’s Data Management Pipeline Builders and CrowdStrike’s CrowdStream (CrowdStream leverages Cribl). Generic DPM tools lack security-specific context. Data pipeline management tools are not new; your enterprise likely uses them already, especially on the data team, but they are likely not specific to the security use case, which makes them more cumbersome for the security team to retrofit to support the security use case. For example, it will become more difficult to transform data to align to a standard like OCSF (Open Cybersecurity Schema Framework), since generic tools will not support the framework. The tools may also lack the integrations into security tools you need. With that said, in upcoming reports, Forrester will be releasing research on data use case crossover and consolidation. In December, I’ll be speaking on security data management strategies at Forrester’s Security & Risk Summit in Baltimore, Maryland. Come join us and get your questions answered! In the meantime, if you have any questions about data pipeline management for security and IT, request an inquiry or guidance session with me or one of my colleagues. source

By Juan-Carlos Rodriguez ( November 15, 2024, 8:39 PM EST) — The U.S. Environmental Protection Agency’s internal watchdog is sounding the alarm about cybersecurity weaknesses in the nation’s drinking water systems, and said there are problems with the plans for reporting and coordinating responses to attacks…. Law360 is on it, so you are, too. A Law360 subscription puts you at the center of fast-moving legal issues, trends and developments so you can act with speed and confidence. Over 200 articles are published daily across more than 60 topics, industries, practice areas and jurisdictions. A Law360 subscription includes features such as Daily newsletters Expert analysis Mobile app Advanced search Judge information Real-time alerts 450K+ searchable archived articles And more! Experience Law360 today with a free 7-day trial. source

As generative AI begins takes hold in business, who does what work and how organizations will be structured will inevitably change, particularly at the leadership and management levels, according to a new survey from Capgemini in which 1,500 managers from 500 organizations and 15 countries participated. The majority (51%) of respondents believe that decision-making positions will become more niche as a result of the use of generative AI. This will mean that leaders will also need to be experts in various areas such as data analysis, AI strategy, ethical assessment, and risk management. As a result, 53% of managers surveyed believe generative AI will shift organization structures to become more diamond-shaped, with fewer junior positions and a larger midlevel management layer. Junior roles are expected to decrease from 44% of the organization today to 32% in three years, while middle managers will increase from 44% to 53%.  source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More SambaNova and Hugging Face launched a new integration today that lets developers deploy ChatGPT-like interfaces with a single button click, reducing deployment time from hours to minutes. For developers interested in trying the service, the process is relatively straightforward. First, visit SambaNova Cloud’s API website and obtain an access token. Then, using Python, enter these three lines of code: import gradio as gr import sambanova_gradio gr.load(“Meta-Llama-3.1-70B-Instruct-8k”, src=sambanova_gradio.registry, accept_token=True).launch() The final step is clicking “Deploy to Hugging Face” and entering the SambaNova token. Within seconds, a fully functional AI chatbot becomes available on Hugging Face’s Spaces platform. The three-line code required to deploy an AI chatbot using SambaNova and Hugging Face’s new integration. The interface includes a “Deploy into Huggingface” button, demonstrating the simplified deployment process. (Credit: SambaNova / Hugging Face) How one-click deployment changes enterprise AI development “This gets an app running in less than a minute versus having to code and deploy a traditional app with an API provider, which might take an hour or more depending on any issues and how familiar you are with API, reading docs, etc…,” Ahsen Khaliq, ML Growth Lead at Gradio, told VentureBeat in an exclusive interview. The integration supports both text-only and multimodal chatbots, capable of processing both text and images. Developers can access powerful models like Llama 3.2-11B-Vision-Instruct through SambaNova’s cloud platform, with performance metrics showing processing speeds of up to 358 tokens per second on unconstrained hardware. Performance metrics reveal enterprise-grade capabilities Traditional chatbot deployment often requires extensive knowledge of APIs, documentation, and deployment protocols. The new system simplifies this process to a single “Deploy to Hugging Face” button, potentially increasing AI deployment across organizations of varying technical expertise. “Sambanova is committed to serve the developer community and make their life as easy as possible,” Kaizhao Liang, senior principal of machine learning at SambaNova Systems, told VentureBeat. “Accessing fast AI inference shouldn’t have any barrier, partnering with Hugging Face Spaces with Gradio allows developers to utilize fast inference for SambaNova cloud with a seamless one-click app deployment experience.” The integration’s performance metrics, particularly for the Llama3 405B model, demonstrate significant capabilities, with benchmarks showing average power usage of 8,411 W for unconstrained racks, suggesting robust performance for enterprise-scale applications. Performance metrics for SambaNova’s Llama3 405B model deployment, showing processing speeds and power consumption across different server configurations. The unconstrained rack demonstrates higher performance capabilities but requires more power than the 9KW configuration. (Credit: SambaNova) Why This Integration Could Reshape Enterprise AI Adoption The timing of this release coincides with growing enterprise demand for AI solutions that can be rapidly deployed and scaled. While tech giants like OpenAI and Anthropic have dominated headlines with their consumer-facing chatbots, SambaNova’s approach targets the developer community directly, providing them with enterprise-grade tools that match the sophistication of leading AI interfaces. To encourage adoption, SambaNova and Hugging Face will host a hackathon in December, offering developers hands-on experience with the new integration. This initiative comes as enterprises increasingly seek ways to implement AI solutions without the traditional overhead of extensive development cycles. For technical decision makers, this development presents a compelling option for rapid AI deployment. The simplified workflow could potentially reduce development costs and accelerate time-to-market for AI-powered features, particularly for organizations looking to implement conversational AI interfaces. But faster deployment brings new challenges. Companies must think harder about how they’ll use AI effectively, what problems they’ll solve, and how they’ll protect user privacy and ensure responsible use. Technical simplicity doesn’t guarantee good implementation. “We’re removing the complexity of deployment,” Liang told VentureBeat, “so developers can focus on what really matters: building tools that solve real problems.” The tools for building AI chatbots are now simple enough for nearly any developer to use. But the harder questions remain uniquely human: What should we build? How will we use it? And most importantly, will it actually help people? Those are the challenges worth solving. source

When B2B marketing leaders start their annual planning, they often lack some critical input. That keeps their plans from being as effective as they could be — and often, it means that marketing activities won’t support business objectives. In my previous post, I explained how the Forrester B2B Marketing Planning Process provides a structure for annual planning efforts. In this post, I will share the three types of information that need to be gathered, distilled, and integrated into a marketing plan. The Business Strategy Is An Input Into Marketing Planning Marketers building a plan need to understand the company’s strategy and how it translates into what the company wants to become, how it plans to get there, and how it will reach buyers. In addition to aligning with the strategy in the plan year, the strategy indicates what needs to be in place in subsequent years, and that may require marketing action in the current year. This information often comes from your C-suite. In an ideal situation, it is part of a documented strategy that has been developed and that coordinates vision, growth, routes-to-market, and operational realities across the sales, marketing, and product functions. But often, this information is not communicated beyond the executive team. And to make it even more challenging, it is aspirational, spans multiple years, and may be too vague to use for planning. Marketing leaders must translate this strategy information into brand, audience, value, and operational terms to know how this will affect marketing priorities, team alignment, marketing annual goals, and resource requirements in the coming year or two. Target Audiences And The Offering Portfolio Are Inputs Into Marketing Planning A marketing plan needs to reflect an understanding of the audience for the company’s offerings and how the offerings will evolve over the plan’s timeframe. If new product offerings or competitive initiatives are in the cards, marketing needs to prepare. Audience and offering information usually can be found by working with your product management and product marketing teams and through your business unit leaders. Ideally, there is a timeline for new product introductions as well as target buyer personas, clarity around buyer needs, and an understanding of the buyer’s journey. Sometimes it can be challenging for marketing to gain access to this information, but clearly conveying how this information will inform near-term and intermediate-term marketing efforts should make it easier. The Business Revenue Plan Is An Input Into Marketing Planning Without an understanding of the revenue plan, marketers are flying blind. It is essential to understand the proportions of revenue that will come from both prospects and existing customers, including retention, upsell, and cross-sell. It’s essential to understand the distribution of planned revenue by route and geography so that the marketing programs that come from the marketing plan are pointed at the right regions and on the right sales teams, partners, and marketplaces. Also important is setting expectations for executive leadership as to what marketing will achieve to support the revenue plan. While we at Forrester try to get our clients to think about measuring marketing engagement that drives desired revenue outcomes, we also know that many companies are still focused on marketing-sourced and -influenced revenue. Without a clear view of the overall revenue plan, committing to contribution levels is a wild guess. The next step of marketing planning is to figure out what to do with all this information. (Hint: You use it to gain alignment on objectives, set an overall approach to interacting with the many audiences you are targeting, set priorities and quantifiable goals, determine action plans, and identify risks.) In my next post, I’ll outline how this information is brought into the B2B Marketing Planning Process and used to develop the B2B-Marketing-Plan-On-A-Page Template (client-only; non-clients can access a version here). source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Alibaba Cloud has released Qwen2.5-Coder, a new AI coding assistant that has already become the second most popular demo on Hugging Face Spaces. Early tests suggest its performance rivals GPT-4o, and it’s available to developers at no cost. The release includes six model variants, from 0.5 billion to 32 billion parameters, making advanced AI coding accessible to developers with different computing resources. This achievement by the Chinese tech company comes despite facing export restrictions on advanced semiconductors. According to the team’s technical report on arXiv, Qwen2.5-Coder’s success stems from refined data processing, synthetic data generation, and balanced training datasets, resulting in strong code generation while maintaining broader capabilities. A comparison of AI coding models shows Alibaba’s Qwen2.5-Coder-32B (in blue) outperforming GPT-4 and other competitors across multiple industry benchmarks. Source: Alibaba Cloud Research State-of-the-art performance raises stakes in global AI race The flagship model, Qwen2.5-Coder-32B-Instruct, has shattered previous benchmarks for open-source coding assistants. It scored 92.7% on HumanEval and 90.2% on MBPP, two crucial metrics for measuring code generation abilities. Most impressively, it achieved 31.4% accuracy on LiveCodeBench, a contemporary benchmark testing AI models on real-world programming challenges. The achievement goes far beyond typical performance metrics. While most AI coding assistants specialize in one or two popular languages like Python or JavaScript, Qwen2.5-Coder’s mastery of 92 programming languages — from mainstream tools to niche languages like Haskell and Racket — represents a major leap forward in AI versatility. This broad language support, combined with its ability to handle complex tasks like repository-level code completion and debugging, suggests we’re entering a new era where AI coding assistants can truly function as universal programming partners rather than just specialized tools. Benchmark results comparing Alibaba’s Qwen2.5-Coder against leading AI models, including GPT-4 and Claude 3.5. The new model (leftmost column) achieves top scores in several key metrics, including a 92.7% accuracy rate on HumanEval, surpassing both open-source and proprietary competitors. Source: Alibaba Cloud Research Open-source strategy could reshape enterprise software development Unlike its closed-source competitors, most Qwen2.5-Coder models carry the permissive Apache 2.0 license, allowing companies to freely integrate them into their products. This could dramatically reduce development costs for businesses worldwide while accelerating AI adoption. The model’s capabilities extend beyond basic coding. It excels at repository-level code completion, understands context across multiple files, and can generate visual applications like websites and data visualizations. “We explore the practicality of Qwen2.5-Coder in two scenarios, including code assistants and Artifacts, with some examples showcasing the potential applications in real-world scenarios,” the researchers explained in their paper. China’s AI innovation defies U.S. chip restrictions This release could fundamentally alter the economics of AI-assisted software development. While companies like OpenAI and Anthropic have built their business models around subscription access to proprietary models, Alibaba’s decision to open-source Qwen2.5-Coder creates a new dynamic. Enterprise customers who currently pay hundreds of thousands of dollars annually for AI coding assistance could soon have access to comparable capabilities at a fraction of the cost. This doesn’t just challenge existing business models – it could accelerate AI adoption among smaller companies and developers in emerging markets who have been priced out of the current AI boom. The shift toward open-source, enterprise-grade AI tools also raises strategic questions for Western tech companies. As more sophisticated open-source alternatives emerge, maintaining high-priced subscription models for AI services may become increasingly difficult to justify to enterprise customers. The achievement is particularly important given the ongoing U.S. restrictions on chip exports to China. Alibaba’s success suggests Chinese tech companies have found ways to innovate despite these constraints, possibly reshaping the global AI competitive landscape. The model’s release intensifies the AI development race between the U.S. and China. While American companies have traditionally led in large language models, Chinese firms are increasingly matching or exceeding their capabilities in specialized domains like coding and mathematics. Alibaba’s researchers plan to explore scaling up both data size and model size while enhancing reasoning capabilities. This suggests the company isn’t content with current achievements and aims to push the boundaries further. For developers and businesses worldwide, Qwen2.5-Coder presents a new option in the AI toolkit — one that combines state-of-the-art performance with the freedom of open-source software. As the AI arms race continues to accelerate, this release may mark a shift in how advanced AI capabilities are distributed and accessed globally. source

I am excited to announce The Forrester Wave™: Point-of-Service Solutions, Q4 2024. When we last evaluated this market in 2018, new cloud point-of-service (POS) solutions offered the safety and excitement of reliability, plus features to engage customers. The goals for POS haven’t changed much in those six years. Then and now, retail firms expect POS to drive omnichannel sales, deliver brand consistency across the empowered consumer’s path to purchase, and boost store associate productivity. What has changed is the internet’s impact on offline retail sales and the demand for seamless omnichannel services. Even in the grocery sector, consumers use digital touchpoints in-store and appreciate convenient checkout and fulfillment options. As a result, more retail firms are assessing their POS technologies and replacing slow or limiting POS solutions that can’t keep up with evolving expectations. Customer references for this Wave mentioned modern architecture, ease of use, and interoperability as key requirements for their new POS. But strategy and partner relationships often sealed the deal. References shared comments such as “We knew we could inform some of the way [the vendor] built [the POS],” “The vendor shared a compelling vision and has delivered on that vision,” and “They keep adding features to meet our needs.” Today’s POS buyers aren’t just looking for a bunch of features to check off their list. They’re also looking for a strategic partner that knows their business and is committed to helping them grow and rapidly adapt to whatever lies ahead. So what do you need to know when selecting a new POS? Vendors differ in how they: Support traditional and emerging checkout experiences in the store. Although most POS vendors provide a responsive UI, they do not equally support fixed, mobile, self, and automated checkout. Some offer comprehensive self-checkout systems with specialized interfaces, kiosk integrations, management tools, etc. Others excel in mobile experiences, utilizing native capabilities such as push notifications for pickup orders. Few have extensive deployments across all touchpoints. Leverage integrations to provide value beyond in-store checkout. A connected and informed POS that easily integrates with adjacent solutions is table stakes. Retail firms expect their POS to not only “see” what’s happening across the business, but it must also expose that data to users in a way that’s maximally useful. This means sleek interfaces and tools that are purpose-built for value-added functions such as clienteling, store fulfillment, and inventory management. Empower nontechnical practitioners to customize the POS experience. Vendors differ in how they equip users with no-code/low-code tools. Some vendors offer sophisticated visual editors that enable nontechnical practitioners to easily adjust the front end, such as modifying the checkout flow, configuring promotional offers, or updating digital receipts. These tools enable quick changes without requiring technical expertise. You can read our full Wave evaluation here and our market overview research on the 2024 POS landscape here. We’ll also host a webinar in early 2025 for Forrester clients about learnings from this evaluation — stay tuned for details. Brands and retailers: Please schedule a guidance session with me to see how to use this research to identify the best-fit solutions for your needs. I’ll walk you through my findings and help you tailor the research to your needs to identify the vendors that should make your shortlist. POS vendors and commerce-related solution providers: Please schedule an inquiry or advisory session with me to discuss what my findings mean for the industry and your offering. source