AI advantage does not come simply by giving employees access to ChatGPT or Microsoft 365 Copilot. These general-purpose models and tools, trained mainly on the world’s public data, cannot differentiate your business or define your long-term success. Nor can buying off-the-shelf AI-accelerated products like Github Copilot or Tabnine to, for example, automate your coding.  Our […] source

By Jade Martinez-Pogue ( January 3, 2025, 3:31 PM EST) — Activist investment firm The D.E. Shaw Group on Friday slammed the board of directors of industrial gas supplier Air Products and Chemicals Inc., accusing it of failing to manage an effective CEO succession process, and called for a change in the board’s composition and the retirement of the company’s current CEO…. 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

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Meta — parent company of Facebook, Instagram, WhatsApp, Threads and more — runs one of the biggest recommendation systems in the world. In two recently released papers, its researchers have revealed how generative models can be used to better understand and respond to user intent.  By looking at recommendations as a generative problem, you can tackle it in new ways that are richer in content and more efficient than classic approaches. This approach can have important uses for any application that requires retrieving documents, products or other kinds of objects. Dense vs generative retrieval The standard approach to creating recommendation systems is to compute, store and retrieve dense representations of documents. For example, to recommend items to users, an application must train a model that can compute embeddings for the users’ requests and embeddings for a large store of items.  At inference time, the recommendation system tries to understand the user’s intent by finding one or more items whose embeddings are similar to the user’s. This approach requires an increasing amount of storage and computation capacity as the number of items grows because every item embedding must be stored and every recommendation operation requires comparing the user embedding against the entire item store. Dense retrieval (source: arXiv) Generative retrieval is a more recent approach that tries to understand user intent and make recommendations not by searching a database but by simply predicting the next item in a sequence of things it knows about a user’s interactions. Here’s how it works: The key to making generative retrieval work is to compute “semantic IDs” (SIDs) which contain the contextual information about each item. Generative retrieval systems like TIGER work in two phases. First, an encoder model is trained to create a unique embedding value for each item based on its description and properties. These embedding values become the SIDs and are stored along with the item.  Generative retrieval (source: arXiv) In the second stage, a transformer model is trained to predict the next SID in an input sequence. The list of input SIDs represents the user’s interactions with past items, and the model’s prediction is the SID of the item to recommend. Generative retrieval reduces the need for storing and searching across individual item embeddings. So its inference and storage costs remain constant as the list of items grows. It also enhances the ability to capture deeper semantic relationships within the data, and provides other benefits of generative models, such as modifying the temperature to adjust the diversity of recommendations.  Advanced generative retrieval Despite its lower storage and inference costs, generative retrieval suffers from some limitations. For example, it tends to overfit to the items it has seen during training, which means it has trouble dealing with items that were added to the catalog after the model was trained. In recommendation systems, this is often referred to as “the cold start problem,” which pertains to users and items that are new and have no interaction history.  To address these shortcomings, Meta has developed a hybrid recommendation system called LIGER, which combines the computational and storage efficiencies of generative retrieval with the robust embedding quality and ranking capabilities of dense retrieval. During training, LIGER uses both similarity score and next-token goals to improve the model’s recommendations. During inference, LIGER selects several candidates based on the generative mechanism and supplements them with a few cold-start items, which are then ranked based on the embeddings of the generated candidates.  LIGER combines generative and dense retrieval (source: arXiv) The researchers note that “the fusion of dense and generative retrieval methods holds tremendous potential for advancing recommendation systems,” and as the models evolve “they will become increasingly practical for real-world applications, enabling more personalized and responsive user experiences.” In a separate paper, the researchers introduce a novel multimodal generative retrieval method named Multimodal preference discerner (Mender), a technique that can enable generative models to pick up implicit preferences from users’ interactions with different items. Mender builds on top of the generative retrieval methods based on SIDs and adds a few components that can enrich recommendations with user preferences. Mender uses a large language model (LLM) to translate user interactions into specific preferences. For example, if the user has praised or complained about a specific item in a review, the model will summarize it into a preference about that product category.  The main recommender model is trained to be conditioned both on the sequence of user interactions and the user preferences when predicting the next semantic ID in the input sequence. This gives the recommender model the ability to generalize and perform in-context learning and to adapt to user preferences without being explicitly trained on them. “Our contributions pave the way for a new class of generative retrieval models that unlock the ability to utilize organic data for steering recommendation via textual user preferences,” the researchers write. Mender recommendation framework (source: arXiv) Implications for enterprise applications The efficiency provided by generative retrieval systems can have important implications for enterprise applications. These advancements translate into immediate practical benefits, including reduced infrastructure costs and faster inference. The technology’s ability to maintain constant storage and inference costs regardless of catalog size makes it particularly valuable for growing businesses. The benefits extend across industries, from ecommerce to enterprise search. Generative retrieval is still in its early stages and we can expect applications and frameworks to emerge as it matures. source

It seems like everyone is talking about performance these days in all aspects of life. High school kids are applying to college, and their SAT/ACT test scores and overall GPAs are being compared to their peers; awards season is kicking off, with nominations for the best performances in TV, movies, and other entertainment areas; and of course, the ever-popular end-of-year business ritual, when employees and managers evaluate their performance over the prior 12 months. It just seems fitting, then, that we kick the year off with a freshly published report on what high performance looks like in portfolio marketing. Results from Forrester’s Portfolio Marketing Survey, 2024, reveal significant gaps between high-performing and low-performing organizations. The survey gathers insights from portfolio and product marketing decision-makers based on their responses to questions about core responsibilities, technology usage, and key activities. High performance is defined as when 80% or more of an organization’s primary offerings are meeting revenue targets, while low-performance organizations are characterized as such by making 40% or less of their offerings’ revenue targets. High-performing portfolio marketers are customer-centric, and our research reveals the characteristics and behaviors that set them apart from their low-performing peers. Existing Forrester clients can access all the survey findings in this full report. Here is just a glimpse at some of the areas that stood out. Audience Mastery Is The Secret Weapon For High Performers One of the most significant responsibilities of a portfolio marketer is to help their organization define and prioritize target market segments and buyer personas. Two-thirds (66%) of high performers indicate that they have defined target market segments for more than half of their offerings, compared to just 19% of low performers, and 62% of high performers have defined target buyer personas for more than half of their offerings (21% for low performers). High Performers Emphasize Market Expertise Understanding market trends and dynamics is a fundamental capability for portfolio marketers. More than half (54%) of high performers own or lead efforts related to market research and intelligence; this is compared to 30% for low performers, and a whopping 35% of low performers say that they don’t do this at all! When it comes to market analysis, 60% of high performers own or lead the effort, versus just 40% for low performers. High-Performing Organizations Make Investments In Portfolio Marketing Portfolio marketing teams have long suffered from being understaffed. As high performers meet or exceed revenue targets, they can support growth within the function. More than half (56%) of high performers indicate that their organization’s investment in the portfolio marketing team will increase by 5% or more over the next 12 months (just 19% for low performers). High performers also invest in ongoing professional development, with 84% dedicating 40 or more hours per year, compared to only 51% of low performers dedicating as much time. For a more in-depth read on the survey findings and how high-performing teams outperform their peers, Forrester clients can check out my recently published report, written with my colleague Nicky Briggs, The State Of Portfolio And Product Marketing In 2024. And if you would like to have a more detailed conversation with me, you can request a guidance session or inquiry. source

There’s a story, probably apocryphal, about a royal leader who presented an elephant as a gift to those who had fallen out of favor. The unlucky recipients were left with a large and expensive problem. They knew they had a valuable animal, and they knew their health depended on the elephant’s health. However, the cost of feeding and caring for the creature exceeded the value they could realize. Ultimately, they went bankrupt, surrounded by piles of… err, tech debt. This story came to mind when I talked with a client recently. Their complaint was that “engineering” and “the business” didn’t understand each other. Engineering was the elephant: bulky, expensive to keep, hard to move, and incomprehensible. Business was the poor beneficiary, not understanding how to create value from the incredible resource — they weren’t even sure what they had. Framing the problem this way is a recipe for failure. Engineering is part of the business, often a large part. Engineering leaders must recognize they need the rest of the business to be healthy, or the elephant will starve. Leaders of the rest of the business must give clear direction to the elephant so it’s doing useful work for the organization, not just wandering off on its own. GenAI: Was That An Earthquake? Developer productivity is on everyone’s mind these days. We’ve all heard that gen AI promises to increase developer productivity by 40% or more. There’s a lot of hype, and leaders need to cut through the hype to find reality. Business leaders both inside and out of engineering have been coming to me with the same question these days: “How can we determine if our developers are more productive with genAI tools?” My response usually doesn’t go over too well: “Take whatever you’re using to measure productivity now, add genAI, and see if those measures go up.” The truth is that measuring developer productivity is hard. Back in 2003, Martin Fowler gave up, saying, “we have no way of reasonably measuring productivity.” Metrics like lines of code have always been meaningless, and they’re even more meaningless when you can add a prompt like, “make this twice as long.” Business Is What Matters In many cases, metrics are a form of vanity. “Our team is DORA elite” doesn’t mean much if your customer doesn’t want what you deliver, or if you’ve got overwhelming turnover costs due to developer burnout. Just the act of measurement will change what happens at your organization, so a light — and balanced — touch is needed. Save the time and motion studies for processes that get repeated and automated. The elephant knows how to lift the log and enjoys doing it. Team up with the elephant so you both succeed. Create alignment — make sure the elephant understands where the log needs to be — clear out the obstacles, and let the elephant figure out how to get it there. To learn how to do that, Forrester clients can connect with me or read my report, “Your Focus On Developer Productivity Is Killing You.” source

While project teams are off to the next initiative, the EPMO measures that the benefits promised in the business case are realized in a way that made them worth the investment. During this phase, the EPMO holds teams accountable, tracks benefits over time, and feeds those insights straight to leadership. This isn’t just reporting, it’s about giving leaders what they need to make smarter decisions and bet on the initiatives that will drive the most value moving forward. Where the EPMO adds value: Benefits and ROI tracking – Measuring long-term performance to ensure projects deliver as expected. Continuous improvement – Running post-project reviews to ensure future initiatives benefit from lessons learned. Operational handover – Transitioning project outputs to operational business owners to achieve the successful business outcomes. Your PMO, whether department-specific or enterprise-focused, can manage the full strategy lifecycle. At the departmental level, a PMO drives initiatives that align with the department’s strategy, while an EPMO ensures alignment across the entire organization. Both play critical roles, but their scope, influence, and impact on organizational culture differ. source

Hong Kong, China – January 02: at Bank of China Hong Kong Tennis Open 2025 ATP250 at Victoria Park Tennis Court on January 2, 2025 in Hong Kong, China. (Photo by Patrick Leung / HKCTA) 中國香港網球總會（「網總」）結合運動與生活的概念，在「中銀香港網球公開賽」推出星級廚房「The Culinary Room」，這不僅廣受本地美食家的好評，亦獲得到不少ATP球員的青睞。同時星級主廚們也十分喜歡這個概念，並把他們的最佳表現帶到比賽中。 今年參與「The Culinary Room」的名廚包括（根據場次先後排列）：米芝蓮一星餐廳MONO的主廚Ricardo Chaneton、Arcane的主廚Shane Osborne、Ryota Keppou Modern的主廚兼澤良太、米芝蓮三星餐廳8½ Otto e Mezzo Bombana的主廚Umberto Bombana，以及Clarence的主廚Olivier Elzer。 中國香港網球總會義務秘書兼香港網球公開賽督導委員會主席周維正表示：「我們很榮幸再次邀請到城中的頂尖名廚及其餐廳團隊，為『中銀香港網球公開賽』的賓客打造此獨特的體驗。在香港的大型體育盛事中，『The Culinary Room』是一個嶄新的概念，中國香港網球總會不斷致力提升賽事體驗，務求讓市民和遊客能夠在假期中享受這種難得的卓越體驗。」 Ryota Keppou Modern的主廚兼澤良太昨天首次亮相於「The Culinary Room」，他融合現代元素的摩登日式料理吸引了不少參賽的頂尖球手，包括世界排名第19、來自世界美食之都法國的費爾斯（Arthur Fils）、賽事大使胡霍特（Mark Woodforde）和卡殊（Pat Cash）一同品嚐晚宴，而2014年美國公開賽亞軍、日本名將錦織圭也被他的美食吸引到場享用了四道菜豐盛季節菜單。 早前，兩位名廚兼澤良太和Umberto Bombana抽空在賽事的「網球同樂區」內輕鬆一下。兩位主廚首次接觸現時全球發展最快的運動之一 —— 匹克球。 網總的匹克球計劃作為網球和拍類運動的入門，深受各年齡層的球迷歡迎。兩位主廚很高興能夠在「中銀香港網球公開賽」內體驗這個新興運動。 賽事金色門票售價由港幣370元至1,610元；銀色門票售價由港幣270元至1,260元。公眾門票於網上售票平台www.tickets.tennishk.org/atp250-2025發售。為響應環保，賽事不設實體門票*。所有銀色門票均設兒童及長者半價優惠**。 「中銀香港網球公開賽」是一項「M」品牌認可活動，協助提升香港作為亞洲體育盛事之都的形象。獲大型體育活動事務委員會頒發的「M」品牌，標誌著緊張、精彩、刺激的大型體育活動。 如欲查詢更多賽事及售票資訊，請瀏覽賽事網頁 www.hkmenstennisopen.com 關於中國香港網球總會 中國香港網球總會是香港的本地體育總會，成立於 1909 年，是亞洲網球聯會及國際網球聯會認可組織。香港網總旗下現時有 40 個屬會和 4,000 多名會員。除了舉辦本地和國際網球比賽，香港網總亦推行青少年和精英發展計劃、工商機構訓練、教練證書課程、普及訓練計劃、和為幼童而設的小型網球訓練等。如欲查詢更多，請瀏覽以下網頁 www.tennishk.org。 官方網站：www.hkmenstennisopen.com 關於ATP巡迴賽                                                                                           作為男子職業網壇的全球管理機構，ATP的使命是為網球運動服務。我們吸引了超過十億全球球迷，在最享負盛名的錦標賽中展現世界上最頂尖球員的巔峰對決，並激發新一代球迷和球員的熱情。從在澳洲舉辦的聯合杯開始，到歐洲、美洲和亞洲等地，這些網壇明星們將會在 ATP 1000大師賽、500和250級別賽事和四大滿貫賽事中爭奪冠軍頭銜和 Pepperstone ATP 排名積分。競逐的道路一致通向意大利都靈舉辦的 Nitto ATP 年終總決賽。年終總決賽僅由世界排名前八名的單打選手和雙打組合參賽，屆時也會見證年終ATP世界第一的誕生，而這個網球界的終極成就將由Pepperstone頒發。欲了解更多詳情，請前往www.ATPTour.com LinkedIn Email Facebook Twitter WhatsApp The post 「中銀香港網球公開賽2025」另類體驗：頂級名廚 x 星級球手 appeared first on VeriMedia. source

On the second Tuesday of each month, Microsoft and other tech companies release patches for consumer and enterprise users. These updates, including bug fixes and security enhancements from the previous month, are known as “Patch Tuesday.” The monthly update is an important opportunity to ensure that security features and applications are up to date. Microsoft details the official Patch Tuesday release in their Security Update Guide. Below, TechRepublic explores its purpose, how it works, and how you can prepare for it. What is the purpose of Patch Tuesday? Patch Tuesday is designed to collect and release many enterprise software updates simultaneously. While some updates are urgent enough to require immediate attention throughout the month, non-urgent or quality-of-life fixes are consolidated for release on Patch Tuesday. “Whether you’re an IT administrator or a general user, Windows monthly updates provide you with the security fixes to help keep your devices protected—as well as enhancements based on your feedback,” wrote Microsoft Senior Director of Communications Chris Morrissey in a 2023 blog post. Patch Tuesday is technically known as Microsoft’s “B” release,” as opposed to “C” and “D” releases, which occur during the third and fourth weeks of the month. Other companies, including Adobe, have followed Microsoft’s lead in rolling out mass patches on the second Tuesday of each month. What to know How does Patch Tuesday work? Administrators and users can access these updates through various tools, including: Windows Update Windows Update for Business Microsoft Intune Microsoft Configuration Manager Windows Server Update Services (WSUS) The Microsoft Update Catalog Before rolling out patches across an organization, administrators should test them in an isolated environment and a small test group. Additionally, administrators should have a rollback plan in place if issues arise. SEE: Microsoft power users may want to watch for monthly PowerToys updates.  Since exploits from the previous month are detailed on Patch Tuesday, the following day often sees a spike in copycat attacks targeting unpatched systems. Organizations should prioritize applying critical security updates to mitigate this risk. As of the February 2023 commercial control update, administrators have some control over which patches to immediately apply. This allows for flexibility in managing updates that introduce new capabilities, remove existing ones, or significantly alter user-facing features, such as the start menu. Microsoft Copilot resources from TechRepublic What is the difference between Patch Tuesday and out-of-band updates? Along with the letter naming system for releases, you might hear the term “out-of-band release” when it comes to patches. Out-of-band releases are not sent out on a schedule like the planned monthly patches. Instead, atypical updates may be sent out at any time to address an ongoing security or quality issue. How can I prepare for Patch Tuesday updates? Admins should have a process for applying Patch Tuesday updates, but these processes will differ based on the organization’s size and needs. Some patches should be applied right away — especially with security fixes for backdoors that are being actively exploited. Admins might want to wait to deploy non-critical patches in case Microsoft issues any revisions. source

The religious composition of the 119th Congress Members of the incoming 119th Congress will be sworn in at the U.S. Capitol on Jan. 3, 2025. (Cynthia Johnson/Getty Images) Every two years, Pew Research Center publishes a report on the religious affiliation of members of the incoming Congress. This report is the ninth in the series, which started with the 111th Congress that began in 2009. Data on members of Congress comes from CQ Roll Call, which surveys members about their demographic characteristics, including religious affiliation. Pew Research Center researchers then code the data so that Congress can be compared with U.S. adults overall. For example, members of Congress who tell CQ Roll Call they are “Southern Baptists” are coded as “Baptists” – a broader category (including Southern Baptists as well as other Baptists) used for analysis of the general public.  Data in this report covers voting members of Congress scheduled to be sworn in on Jan. 3, 2025. While there are 535 voting seats in Congress (100 in the Senate and 435 in the House of Representatives), this analysis excludes three of those seats: Florida’s 1st District seat, due to the resignation of Matt Gaetz; Florida’s 6th District seat, due to the announced resignation of Michael Waltz; and the Ohio Senate seat held by JD Vance, who is set to become vice president on Jan. 20, 2025. This analysis, then, looks at 532 members of Congress rather than 535. Data for all U.S. adults comes from Pew Research Center’s 2023 National Public Opinion Reference Survey (NPORS), conducted May 19-Sept. 5, 2023. Figures for Protestant subgroups, Messianic Jews, Unitarians and Humanists come from the Center’s American Trends Panel (ATP) survey conducted Aug. 7-27, 2023. Jewish estimates come from the Center’s survey of Jewish Americans, conducted Nov. 19, 2019-June 3, 2020. Read more about how Pew Research Center measures the religious composition of the United States. When the U.S. Congress convenes for its 119th session on Jan. 3, it will have marginally fewer Christians than it did in the previous session (2023-25), continuing a gradual, 10-year decline. Christians will make up 87% of voting members in the Senate and House of Representatives, combined, in the 2025-27 congressional session. That’s down from 88% in the last session and 92% a decade ago. Overall, there will be 461 Christian members of Congress when the 119th Congress meets, compared with 469 in the previous Congress and 491 during the 2015-17 session. It will be the lowest number of Christians since the start of the 2009-2011 congressional session, the first for which Pew Research Center conducted this analysis. (This analysis does not include three vacant – or soon to be vacant – seats whose eventual occupants are unknown, including the Ohio Senate seat of Vice President-elect JD Vance.) And yet, at 87%, Christians still make up the lion’s share of the Congress, far exceeding the Christian share of all U.S. adults, which stands at 62% after several decades of decline. In 2007, 78% of American adults were Christian, according to Pew Research Center’s Religious Landscape Study from that year, and in the early 1960s more than nine-in-ten U.S. adults were Christian, according to historical Gallup polling. The new Congress is also more religious than the general population by another, related measure: Nearly three-in-ten Americans (28%) are religiously unaffiliated, meaning they are atheist or agnostic or say their religion is “nothing in particular.” But less than 1% of Congress falls into this category, with three religiously unaffiliated members: incoming Reps. Yassamin Ansari of Arizona and Emily Randall of Washington, both of whom are Democrats, and incoming Rep. Abraham Hamadeh of Arizona, a Republican. While the share of the U.S. public that is religiously unaffiliated – sometimes called “nones” – has risen rapidly in recent decades (from 16% in 2007 to 28% in our recent polling), the corresponding share of Congress has remained miniscule. Prior to the 119th session, the only member of Congress who was categorized as religiously unaffiliated in our analyses was Kyrsten Sinema, independent of Arizona, who served from 2013 through the Congress that is just ending. (She did not run for reelection in 2024.) Pew Research Center’s analysis is based on data from CQ Roll Call, a publisher in Washington, D.C., that has closely covered Congress for decades. Breakdown by denomination Of the 461 Christians in the 119th Congress, 295 are Protestant, a decrease of eight from the previous session. Partial historical data suggests that Protestants had a much larger presence in Congress a few decades ago, including 398 members in 1961. But there have been fewer than 300 Protestants in six of the last nine sessions over the last decade and a half. That said, Protestants continue to make up a disproportionately high share of the 119th Congress (55% of members) when compared with the U.S. adult population (40%). Baptists are the largest category of Protestants in the new Congress, with 75 members (14.1% of Congress). That’s eight more Baptists than in the prior session. The next largest Protestant groups in the new Congress are Methodists (26 members), Presbyterians (26), Episcopalians (22) and Lutherans (19). These four groups have had shrinking U.S. memberships in recent decades and now have a considerably smaller presence in Congress than they used to. For example, in the 112th Congress of 2011-13, there were 51 Methodists, 45 Presbyterians, 41 Episcopalians and 26 Lutherans. Of the 295 Protestants in Congress, 101 do not specify a particular denomination or denominational family, instead giving broad or vague answers such as “Protestant,” “Christian” or “evangelical Protestant.” This is six fewer who identify in those ways than in the last Congress, but the overall trend during the last decade has been for increasing numbers of U.S. representatives and senators to give these kinds of answers. By comparison, only 58 members said they were “just Christians” or gave nonspecific, Protestant descriptions of their religious affiliation at the start of the 114th Congress in 2015. The new Congress also has 150 Catholics, two more than in