Every week, I get inquiries about technical debt, whether from enterprise architects, CIOs, CTOs, offices of the CIO, or strategic portfolio managers. It’s a broad concern in IT and digital management, as well as in the minds and operating models of Forrester clients, that has expanded well beyond the original Ward Cunningham definition. Forrester’s current definition, based on our client conversations and survey data, is as follows: Deferred IT investment to address tech sprawl, aging hardware, obsolete technology, security vulnerabilities, disparate data, hastily written software, legacy processes, and other IT concerns that increase costs, reduce resilience, and blunt business outcomes Purists holding to the original definition will disagree, yet in Forrester’s Modern Technology Operations Survey, 2024, we asked respondents, including developers with exposure to day-two concerns, what the concept means to them. The leading responses included: Out-of-date skills. Obsolete hardware and software. Redundant IT systems. “Poorly or hastily written code,” over the last two years, was the least frequently chosen response, meanwhile. We examine this data, along with various industry cases, and present guidance in the new report, The Forrester Guide To Technical Debt. There’s little question that technical debt can reach crisis proportions in the modern digital enterprise. This year, I hear a new phrase on the lips of Forrester clients: technical bankruptcy. At this point, business outcomes are impacted — new system initiatives encounter constant schedule and cost overruns, existing systems are redundant and poorly resilient, and security risks escalate past an acceptable level. How do enterprises get to this point? This year, I did some economic modeling of the question with a large client, and what became clear is that technical debt is “death by a thousand cuts.” Just like an out-of-control credit card, it happens one Starbucks purchase at a time; any individual decision seems marginal and defensible, but the aggregate impact is ultimately dire. Digging further in, one scenario we hear too often is “We know that we have huge technical debt, but we can’t make the business case to remediate it, as it doesn’t have the necessary ROI.” I’ve heard this both from clients and from the executives of global systems integrators trying to help customers make the business case for such programs (as a GSI offering). This makes me crazy. Think about it. Technical debt is a function of already-acquired digital/IT assets. To equate it to another well-known kind of debt: Do you ask “What’s the ROI?” on your monthly car payment? Do you ask “What’s the ROI?” on your car’s necessary maintenance? Do you ask “What’s the ROI?” when your poor old car is broken down, parked in front of your house, and you need to pay to have it junked ASAP? (Yeah, there’s no salvage value in many old IT systems.) Of course you don’t. You think about ROI up front and accept all of these costs and prepare for them if you are a responsible acquirer of an asset. And as an IT leader, you are not managing one personal vehicle. You are managing a fleet. Do you think Hertz is surprised by maintenance or disposal costs and nickeling-and-diming every last car? Part of the problem here is the ongoing legacy of the old waterfall model — as an industry, we are still struggling with its assumptions. A continuously evolving product requiring ongoing refactoring of tech debt is like having a monthly car payment. But in the waterfall model, we assumed that the vehicle was completely “paid” on project conclusion and threw it over the wall into operational mode, where the assumption was that it needed minimal further investment — just “run” support from shared IT services. While some (mainly packaged) SW still fits this model, digital organizations have long since moved past it (via “project to product”) for their most critical systems. The big gap to get the modern IT organization to a more rational, mature model for managing technical debt is operationalization. Big-batch programs are expensive, run into the ROI issue, and (too often) lose steam and don’t deliver on their expectations. You need to understand and calculate your technical debt (and unfortunately, there is no standard industry metric because of the breadth of the concept.) You need to have an architecture in place to track and report on technical debt that is ongoing. Technical debt needs to be understood as being made up of actionable cases, and those cases need to be routed to your project, product, and/or capability teams on equal standing with the shiny new features that your stakeholders want (hat tip to Mik Kersten’s Flow Framework.) This understanding and financial approach means that you need to dedicate a percentage of portfolio funding to debt reduction. Forrester recommends (in agreement with Marty Cagan of the Silicon Valley Product Group) that this number start at 20% of your portfolio spend. “Peanut butter” opex under constant pressure to cut, cut, cut will not get you there. Have a look at our report and/or ping me on LinkedIn. *(As yet unpublished — Forrester clients, I can share sanitized, high-level model findings on inquiry.) source

Social media plays a pivotal role in shaping brand strategies and reaching target audiences for B2B companies. Forrester’s 2024 B2B Brand And Communications Survey sheds light on the preferences and strategies of over 100 marketing leaders from B2B companies with revenues exceeding $100 million. Our research reveals a clear leader in the realm of social media marketing: LinkedIn. LinkedIn — The Unrivaled Leader LinkedIn, designed specifically for professional networking, stands out as the top choice for B2B marketers. The platform provides unique access to B2B buyers, making it the preferred social media channel by a significant margin over other platforms. This dominance is reflected in the survey results, where almost all respondents reported that their companies maintain an official branded handle on LinkedIn. LinkedIn far outranks other platforms, with 87% indicating they have a paid relationship with the platform. While LinkedIn leads the pack, other long-standing platforms such as YouTube, Facebook, and Instagram are also part of the B2B marketing mix — more than 50% of respondents reported having official branded handles on these platforms. When it comes to paid relationships, LinkedIn far outranks other platforms, with 87% of survey respondents indicating they have a paid relationship with the platform. This is more than twice the number of those who have paid relationships with the second-ranked platform, Facebook, and an even greater disparity shows with YouTube and Instagram. While these platforms are valuable for organic content, LinkedIn is the go-to for paid B2B marketing efforts. The Decline Of X And The Rise Of Emerging Platforms X (formerly Twitter) has seen a decline in priority among B2B marketers, largely due to concerns around brand safety, content moderation policies, and controversies under Elon Musk’s leadership. Despite this, 59% of marketers still keep X on their radar and maintain an official branded handle for their company. Emerging platforms like TikTok and Meta’s Threads, as well as forums such as Reddit, are gaining traction, particularly among companies targeting Gen Zers and Millennials. These platforms remain experimental options for most B2B companies, however. The presence of branded handles on these platforms is limited, and paid relationships are even rarer, reflecting their current status as niche channels rather than mainstream B2B marketing tools.   Being Social Means Being On Multiple Platforms Maintaining a branded handle on social media platforms involves significant investment in creative content, strategy, operations, and governance. Nevertheless, nearly three-quarters of companies manage branded handles on four or more platforms, highlighting the importance of a multiplatform presence. For B2B marketing leaders, the insights from Forrester’s 2024 survey underscore the importance that LinkedIn plays in their social media strategies. While other platforms such as YouTube, Facebook, and Instagram play supportive roles, LinkedIn’s unique access to B2B buyers and its professional networking capabilities make it the unrivaled leader in B2B social media marketing. Forrester clients can find the full results in our report, The State Of B2B Social Media Marketing Strategy And Preferences, 2024. View additional research by Karen Tran or schedule a call with us today. source

Companies experimented boldly with generative AI and other emerging technologies in 2024. Now, the focus is shifting. 2025 will be about the pursuit of near-term, bottom-line gains while competing for declining consumer loyalty and digital-first business buyers. Some leaders will pursue that goal strategically, in ways that set up their organizations for long-term success. Others won’t — and will come up against the limits of quick fixes. Savvy leaders will use expected budget increases to shore up fundamentals, bolstering infrastructure, streamlining operations, and upskilling employees. As they look to operationalize lessons learned through experimentation, they will deliver short-term wins and successfully play the genAI — and other emerging tech — long game. Strengthening foundations will serve companies well as they navigate looming unknowns, from the outcome of the US presidential election to early enforcement of the EU AI Act. Our Predictions 2025 reports, which publish this week, delve into the forces that will define the business landscape next year. (Clients can access the reports and other Predictions resources here.) The reports shed light on how genAI will evolve and what to expect from a tighter regulatory climate. They explore what’s ahead for private cloud, customer experience, and shifting business buying dynamics (and much more). They explain how trust will be won and lost across global regions. They will help you navigate 2025’s unique dynamics to come out ahead. If you’re not yet a Forrester client, you can access our Predictions hub with complimentary resources, including guides for technology and security, B2B marketing and sales, and B2C marketing and customer experience leaders. You can also register for our companion webinars, which will include live Q&As with the analysts behind our predictions. The year ahead will be pivotal in positioning your company to win in a future increasingly shaped by AI. Use our predictions to help light the way. source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More The number of platforms being released to help enterprises integrate AI agents into their technology stack is not slowing down as the year winds down.  Work management platform Asana already has an AI agent service, but with its new AI Studio feature, the company wants its customers to think about AI agents as part of their larger workflow. AI Studio lets users build workflows on Asana and then deploy multiple custom AI agents directly on the workflow. Customers can create the agents without code and allow them to “take on the busywor,” including handling project coordination.  “The difference here is that we’ve opened up the toolkit to create agents to the folks who build workflows at companies, the folks that orchestrate a large body of work,” said Alex Hood, Asana’s chief product officer, in an interview with VentureBeat. “We can now bring agents to all the places where teams show up to hand off work, inserting an AI agent to take work off people’s plates.”  Hood cited a recent Asana 2024 State of Work Innovation Report that showed 53% of an employee’s time is spent on “busy work,” with unproductive meetings doubling since 2019. By bringing in agents, Hood said, teams are freed from doing tasks crucial to a workflow, such as intake for some marketing teams, to focus on other important work.  AI Studio, which will be a tool integrated into Asana but with an additional fee for access, is built on Asana’s Work Graph data mode, which tracks cross-functional work in an organization.  “There are other platforms who are building AI agents and are doing it on top of places where they have specialties,” Hood said. “For us, our specialty is the Work Graph, the place where work happens and the workflows powering that work leverages the right data.” Customers saw an improvement One of Asana’s first customers to use AI Studio is the financial data company Morningstar.  Hood said Morningstar used AI Studio to centralize IT project requests to streamline the workflow to evaluate new projects. Belinda Hardman, director of Program Management at Morningstar, said in a press release that the new workflow helped the company “eliminate time spent on manual back-and-forth because Asana AI identifies and captures the information we need right off the bat.” Islands of AI agents Agents have become the hot topic in AI this year, with several companies announcing either a platform to customize agents or to access a library of ready-made agents.  To name a few: Microsoft announced it will release a suite of AI agents for its Dynamics 365 service this week. Salesforce released Agentforce last month, and ServiceNow launched its agent library on its Now Assist platform. Asana’s earlier released agentic system joins Agentforce, and other ready-made agents from other service providers will be integrated into Slack.  “The things we might have dreamed of and were talking about a couple of years ago are playing out now because the models are getting that much better,” Hood said. “But the models can’t create great agents on their own. They need to be hooked into software and we, as builders, have gotten good at figuring out how to best integrate deeply AI capabilities.” However, many of these agents — even those embedded in third-party applications like Slack — still function as individual islands of agents talking to other agents built on the same platform. The next frontier for agents coming from workflow systems or other enterprise-focused software will be the ability to communicate with other agents elsewhere.  We’re not there yet, but as more enterprises become comfortable with AI agents and begin deploying these into their organizations, that future may come soon enough.  source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More SambaNova Systems and Gradio have unveiled a new integration that allows developers to access one of the fastest AI inference platforms with just a few lines of code. This partnership aims to make high-performance AI models more accessible and speed up the adoption of artificial intelligence among developers and businesses. “This integration makes it easy for developers to copy code from the SambaNova playground and get a Gradio web app running in minutes with just a few lines of code,” Ahsen Khaliq, ML Growth Lead at Gradio, said in an interview with VentureBeat. “Powered by SambaNova Cloud for super-fast inference, this means a great user experience for developers and end-users alike.” The SambaNova-Gradio integration enables users to create web applications powered by SambaNova’s high-speed AI models using Gradio’s gr.load() function. Developers can now quickly generate a chat interface connected to SambaNova’s models, making it easier to work with advanced AI systems. A snippet of Python code demonstrates the simplicity of integrating SambaNova’s AI models with Gradio’s user interface. Just a few lines are needed to launch a powerful language model, underscoring the partnership’s goal of making advanced AI more accessible to developers. (Credit: SambaNova Systems) Beyond GPUs: The rise of dataflow architecture in AI processing SambaNova, a Silicon Valley startup backed by SoftBank and BlackRock, has been making waves in the AI hardware space with its dataflow architecture chips. These chips are designed to outperform traditional GPUs for AI workloads, with the company claiming to offer the “world’s fastest AI inference service.” SambaNova’s platform can run Meta’s Llama 3.1 405B model at 132 tokens per second at full precision, a speed that is particularly crucial for enterprises looking to deploy AI at scale. This development comes as the AI infrastructure market heats up, with startups like SambaNova, Groq, and Cerebras challenging Nvidia’s dominance in AI chips. These new entrants are focusing on inference — the production stage of AI where models generate outputs based on their training — which is expected to become a larger market than model training. SambaNova’s AI chips show 3-5 times better energy efficiency than Nvidia’s H100 GPU when running large language models, according to the company’s data. (Credit: SambaNova Systems) From code to cloud: The simplification of AI application development For developers, the SambaNova-Gradio integration offers a frictionless entry point to experiment with high-performance AI. Users can access SambaNova’s free tier to wrap any supported model into a web app and host it themselves within minutes. This ease of use mirrors recent industry trends aimed at simplifying AI application development. The integration currently supports Meta’s Llama 3.1 family of models, including the massive 405B parameter version. SambaNova claims to be the only provider running this model at full 16-bit precision at high speeds, a level of fidelity that could be particularly attractive for applications requiring high accuracy, such as in healthcare or financial services. The hidden costs of AI: Navigating speed, scale, and sustainability While the integration makes high-performance AI more accessible, questions remain about the long-term effects of the ongoing AI chip competition. As companies race to offer faster processing speeds, concerns about energy use, scalability, and environmental impact grow. The focus on raw performance metrics like tokens per second, while important, may overshadow other crucial factors in AI deployment. As enterprises integrate AI into their operations, they will need to balance speed with sustainability, considering the total cost of ownership, including energy consumption and cooling requirements. Additionally, the software ecosystem supporting these new AI chips will significantly influence their adoption. Although SambaNova and others offer powerful hardware, Nvidia’s CUDA ecosystem maintains an edge with its wide range of optimized libraries and tools that many AI developers already know well. As the AI infrastructure market continues to evolve, collaborations like the SambaNova-Gradio integration may become increasingly common. These partnerships have the potential to foster innovation and competition in a field that promises to transform industries across the board. However, the true test will be in how these technologies translate into real-world applications and whether they can deliver on the promise of more accessible, efficient, and powerful AI for all. source

道指漲161.35點，漲幅為0.37%，報43239.05點；納指漲6.53點，漲幅為0.04%，報18373.61點；標普500指數跌1.00點，跌幅為0.02%，報5841.47點。 道指再創歷史新高。英偉達領漲芯片股。美國9月零售銷售略超預期，凸顯消費繼續支撐經濟，強化了美聯儲降息25個基點的理由。歐洲央行第三次降息。 歐洲央行官員據報認同12月減息可能性大 通脹接近2%目標。 國際金價創歷史新高，COMEX黃金期貨主力合約上漲16.20美元/盎司，漲幅0.60%，報2707.50美元/盎司。 深圳公積金租房提取階段性政策續期一年。 西安推優化房市政策：全市商品住房取得不動產權證書即可上市交易。 據報內地多家大行今起再下調存款利率 有大行定期存款掛牌利率下調25個基點。 軟銀集團將阿里巴巴持股比例下降至10.93%。 兗煤澳大利亞(3668)第三季度權益煤炭銷量為1040萬噸，環比增長21%，同比增長20%。 兗礦能源(1171)：前三季度商品煤銷量10134萬噸，同比增長2.18%。其中三季度商品煤銷量3346萬噸，同比增長0.68%。 六福集團(590)第二季度整體零售值按年下跌16%。 太平洋航運(2343)第三季度超靈便型干散貨船的現貨市場日均租金水平同比增加45%。 福耀玻璃(3606)發佈前三季度業績 歸母淨利潤54.79億元 同比增加32.79%。 新華保險(1336)：前9月累計原保險保費收入為1456.44億元，同比增加1.91%。 LinkedIn Email Facebook Twitter WhatsApp source

Screenshot 面對環球經濟及地緣政治持續不穩定的挑戰，一年一度的「數碼港創業投資論壇」(CVCF）今年於 10 月24 日至 10 月 25 日以「線上+線下」混合形式舉行。論壇將匯聚來自全球具影響力的投資者和企業家，共同探討各地創投趨勢、Web3.0 及人工智能等新興產業助初創企業與投資者突破界限，把握投資機遇。 數碼港作為香港數碼科技旗艦，致力協助本地初創企業於國際市場發揮香港的創科實力，同 時推動國際創投格局發展。近年，香港政府推出一連串措施，包括最新宣佈設立 100 億元 「創科產業引導基金」，推出人工智能資助計劃，以至加強與中東地區在創新科技上的合 作，大力推動創科生態圈多元發展。「數碼港創業投資論壇」作為創投市場的年度盛事，擔 當至關重要的角色，雲集創投精英，聚焦本地及全球創投成功案例以及數碼科技領域創投市 場的發展趨勢，建構平台對接初創企業家與投資者，促成政產學研投用的交流合作，推動本地及國際夥伴合作以促進科技產業發展。 論壇首日將邀請香港特別行政區政府創新科技及工業局局長孫東教授致開幕辭，並邀得畢馬 威會計師事務所、海闊天空創投、Jungle Ventures、1955 Capital、Draper Dragon、 HALA Ventures、SparkLabs Group 、星際創氪、Intudo Ventures、Global Ventures 等創投界精英出席。 數碼港首席公眾使命官陳思源表示:「數碼港作為香港數碼科技旗艦及創業培育基地，一直致力提升初創企業的融資能力，加速孵化企業成長。『數碼港創業投資論壇』將匯聚環球菁英、投資者和企業家，共同探討全球資本市場，重點聚焦潛力無限的中東、東盟、亞洲及中國内地市場，以及 Web3.0、人工智能等嶄新智能科技產業，提供機會讓初創企業展現創新成果，促進更多融資配對，引領業界深入討論並發掘創新科技，推動創科社群迎接各項挑戰，把握發展機遇、取得突破。」 數碼港投資者網絡策劃小組主席、中手游科技集團有限公司共同創辦人及副董事長、國宏嘉 信資本創始合夥人冼漢廸表示:「儘管全球經濟和地緣政治存在不確定性，在國家《十四五 規劃》框架的支持下，配合香港特區政府《香港創新科技發展藍圖》以及今年財政預算案中對人工智能資助計劃、科技研究支援及新型工業加速計劃的措施，數碼港將繼續透過『數碼 港投資者網絡』及『數碼港投資創業基金』等項目連結企業家和投資者，擴大與亞洲以至東 盟地區的獨特聯繫，發揮國際市場優勢，為香港創科及創投市場發展引入動力，推動香港全 方位建設國際創新科技中心。」 環球市況疲弱  數碼港融資表現出色 面對波動及不穩定的創投市場，數碼港社群初創在過去一年仍有強勁的融資表現。2023 年 10 月至 2024 年 9 月期間，數碼港初創成功籌集了超過 37 億港元的資金，按年增加 23%， 近期完成高額融資的數碼港企業包括 KLOOK、Leapstack、Buy&ship、MediConCen、 DeBox 等，累計融資總額突破 412 億港元。年内，香港第二間持牌虛擬資產交易平台 HashKey Group 晉身成為數碼港第八間獨角獸企業，反映Web3.0 科技的應用潛力吸引投資者支持，以及數碼港推動發展 Web3.0 產業續見成效。 適逢今年「數碼港投資者網絡」(CIN)成立七周年，一直保持強勁增長，透過擴展本地及海内外的投資者網絡，對接初創項目及投資者，持續提升初創企業融資能力。網絡自 2017 年成立至今，總投資額高近26億港元，按年增加 64%。網絡累計促成 96 個項目對接，亦比去年多出 21 個項目，按年增加 1.5 倍。而投資單位亦比去年增加 30 多個，總數超過 200 個投資單位，其中 15%來自包括大灣區，而 14%來自亞太地區，可見網絡成員的國際化背景，匯聚環球創投資金。藉著網絡的豐富基礎，數碼港早前更推出「Web3.0 投資圈」，為高潛力、高增長的 Web3.0 初創企業配對投資者，提高融資成功機會，進一步推動 Web3.0 生態圈發展。 數碼港亦一直透過「數碼港投資創業基金」(CMF)投資高潛力初創企業並為其引資，提升 初創企業的引資能力。截至 2024 年 9 月，基金已經投資 28 個初創企業項目，基金投資連 同共同投資總額超過 19.4億港元，引資比率為 1:9，可見數碼港顯著的引資實力。 更多詳情請瀏覽 http://cvcf.cyberport.hk/ (相) 數碼港首席公眾使命官陳思源預告「2024 數碼港創業投資論壇」活動重點，並介紹今年首辦，將於 10 月 25 日舉行的「Web3.0 創新博覽」，屆時將會展示 Web3.0 技術在各行各業中的實際應和成功案例。 LinkedIn Email Facebook Twitter WhatsApp source

As 2025 approaches, technology leaders face a pressing question: “Where’s the return on investment for our existing IT spend?” Over recent years, companies have heavily invested in cloud, SaaS, and AI solutions, yet many are still struggling to capture their full potential. This challenge is fueling a more cautious approach to budgeting, with a focus on cost efficiency and maximizing the value of past investments. Despite this caution, optimism persists: According to a Forrester survey, 91% of global technology decision-makers plan to increase IT spending, with over half expecting growth to surpass 5% — outpacing inflation. Forrester forecasts that the digital economy will grow at a 6.9% CAGR from 2023 to 2028, and AI enthusiasm remains strong, albeit tempered by a shift to pragmatic delivery. Naturally, these increases come with high expectations for enhanced operational value. The emphasis should be made now to make existing technologies work harder through improved adoption, integration, and optimization. This climate underscores the urgent need for high-performance IT — a strategy that focuses on continuously improving business results though technology and advocates for a balanced approach to investments, divestments, and innovations. To help tech leaders navigate these challenges, here is a look at three of the predictions that we think show what is in store for technology leaders in 2025: Only one in five tech execs on the hook for digital transformation will succeed. Many companies are committing to large-scale transformations, but delivery is painfully slow. Financial results from service giants like Accenture and Capgemini show that bookings for large deals are up, yet many digital transformations stall due to the complexity of coordinating with business, operations, HR, and IT leaders. Successful tech leaders will need to align closely with business peers, adapt quickly to changing market dynamics, and consider switching to co-innovation partners for better value orchestration. Seventy percent of IT organizations will incorrectly turn their back on early career development. The demand for highly specialized software developers with AI experience has surged, creating a split in the workforce. Entry-level positions have dwindled, reducing opportunities for early-career developers and midlevel managers. Over time, this is a recipe for disaster, as this trend threatens the talent pipeline critical for building a high-performing delivery organization. Tech executives need to focus more on building an organization with continuous skills and career development, rather than relying on the market to have readily available talent. In the wake of generative AI disappointments, 25% of tech execs will make employee experience the killer app. Generative AI will improve labor productivity only after firms redefine workflows and drive near-universal adoption. Successful tech execs will prioritize tools that make employees’ lives easier, starting with those under their control, like GitHub Copilot and chatbot IT helpdesks. They will expand to include Microsoft 365 Copilot deployments, ensuring that new tools fit seamlessly into employees’ work processes through human-centered design practices. Forrester clients can read our full Predictions 2025: Tech Leadership report to get more detail about each of these predictions, plus two more bonus predictions. Set up a Forrester guidance session to discuss these predictions or plan out your 2025 technology strategy. If you aren’t yet a Forrester client, you can learn how to put these predictions into action during our live webinar. You can also download our complimentary Predictions guide, which covers our top technology and security predictions for 2025. Get additional complimentary resources, including webinars, on the Predictions 2025 hub. source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Stability AI is out today with a major update for its text to image generative AI technology with the debut of Stable Diffusion 3.5. A key goal for the new update is raise the bar and improve upon Stability AI’s last major update, which the company admitted didn’t live up to its own standards. Stable Diffusion 3 was first previewed back in February and the first open model version became generally available in June with the debut of Stable Diffusion 3 Medium. While Stability AI was an early pioneer in the text to image generative AI space, it has increasingly faced stiff competition from numerous rivals including Black Forest Labs’ Flux Pro, OpenAI’s Dall-E, Ideogram and Midjourney. With Stable Diffusion 3.5, Stability AI is looking to reclaim its leadership position. The new models are highly customizable and can generate a wide range of different styles. The new update introduces multiple model variants, each designed to cater to different user needs.Stable Diffusion 3.5 Large is an 8 billion parameter model that offers the highest quality and prompt adherence in the series. Stable Diffusion 3.5 Large Turbo is a distilled version of the large model, providing faster image generation. Rounding out the new models is Stable Diffusion 3.5 Medium, which has 2.6 billion parameters and is optimized for edge computing deployments. All three of the new Stable Diffusion 3.5 models are available under the Stability AI Community License, which is an open license that enables free non-commercial usage and free commercial usage for entities with annual revenue under $1 million. Stability AI has an enterprise license for larger deployments. The models are available via Stability AI’s API as well as Hugging Face. The original release of Stable Diffusion 3 Medium in June, was a less than ideal release. The lessons learned from that experience have helped to inform and improve the new Stable Diffusion 3.5 updates. “We identified that several model and dataset choices that we made for the Stable Diffusion Large 8B model were not optimal for the smaller-sized Medium model,” Hanno Basse, CTO of Stability AI told VentureBeat. “We did thorough analysis of these bottlenecks and innovated further on our architecture and training protocols on the Medium model to provide a better balance between the model size and the output quality.”  How Stability AI is improving text to image generative AI with Stable Diffusion 3.5 As part of building out Stable Diffusion 3.5, Stability AI took advantage of a number of novel techniques to improve quality and performance. A notable addition to Stable Diffusion 3.5 is the integration of Query-Key Normalization into the transformer blocks. This technique facilitates easier fine-tuning and further development of the models by end-users. Query-Key Normalization makes the model more stable for training and fine-tuning. “While we have experimented with QK-normalization in the past, this is our first model release with this normalization,” Basse explained. “It made sense to use it for this new model as we prioritized customization.” Stability AI has also enhanced its Multimodal Diffusion Transformer MMDiT-X architecture, specifically for the medium model. Stability AI first highlighted the MMDiT architecture approach in April, when the Stable Diffusion 3 API became available. MMDiT is noteworthy as it blends diffusion model techniques with transformer model techniques. With the updates as part of Stable Diffusion 3.5, MMDiT-X is now able to help improve image quality as well enhancing multi-resolution generation capabilities Prompt adherence makes Stable Diffusion 3.5 even more powerful Stability AI reports that Stable Diffusion 3.5 Large demonstrates superior prompt adherence compared to other models in the market.  The promise of better prompt adherence is all about the models ability to accurately interpret and render user prompts. “This is achieved with a combination of different things – better dataset curation, captioning and additional innovation in training protocols,” Basse said. Customization will get even better with ControlNets Looking forward, Stability AI is planning on releasing a ControlNets capability for Stable Diffusion 3.5.  The promise of ControlNets is more control for various professional use cases. StabilityAI first introduced ControlNet technology as part of its SDXL 1.0 release in July 2023. “ControlNets give spatial control over different professional applications where users, for example, may want to upscale an image while maintaining the overall colors or create an image that follows a specific depth pattern,” Basse said. source