Challenges Abound in Scaling AI
I’m a podcaster and an avid podcast listener. There are so many to choose from, but I typically subscribe to a handful of podcasts I listen to regularly. Five days ago, I discovered Silo AI’s (a Finland-based AI developer) podcast, Inference. It discussed scaling AI, a topical question also in the construction sector.
The discussion between CTO Niko Vuokko and COO Jaakko Vainio touched on many challenges that anyone trying to introduce and implement AI in their business will face.
Drawing parallels from aviation
Sometimes, the current state of AI is likened to that of the aviation industry 100 years ago. Back then, basic technology existed, but it took a lot of effort and vast sums of money to get to where we are today.
The time it takes for AI to mature might be much shorter, but many things regarding rules and regulations, infrastructures, safety, and the whole ecosystem will have to be developed.
Developing AI is also in the early stages. There are tools and workflows for developing traditional software. If you have a bug, you have a straightforward process to fix it. In AI, data is driving the logic and behavior of the product. A bug often means fixing the data set. That’s not easy if you have ten terabytes of data. The tools for doing that are still rare.
A tough sell
Silo AI’s latest Nordic State of AI report highlights the issue hampering AI development: the lack of talent—the percentage of survey respondents who mention that as the number one problem has been growing every year.
Even if the company had the resources to invest in AI development, the classical innovator’s dilemma would persist. A company knows that investing in improving its existing products and services will satisfy the current customers’ requirements and even improve productivity and satisfaction.
Why allocate funds to something the company knows little about and is still determining the results? It is a tough sell.
“Well, we're going to invest in this thing that is really uncertain. We don't know if it's going to work, and we really have no clue what the final business value is there, because it doesn't really exist yet. We're just expecting that this market is going to exist in the future.”
As a result, customers invest in AI experiments that don’t hurt anyone if they fail. On the other hand, even if the experiment succeeds, it does not deliver much value or solve a critical problem. Also, experiments and AI teams or labs can be scattered throughout the organization without coherent vision or coordination.
Decision-making is complex but necessary
The podcast discussed the situation during the early days of the Internet. Some companies realized that the new technology could transform their business or create a new business model. Others used the internet to publish a home page with a “contact us” button. The latter ones did not flourish in the long run.
The conversation also touched on the differences between countries regarding AI implementation. The discussion was about something other than the construction sector but businesses in general.
In Finland, companies are considering setting up a committee to plan what to do with AI. In Sweden, the committee is charting the plan down. In the Netherlands and Germany, companies have started planning the execution. It seems that US companies have executed the plan for a while.
Integration is the key
AI as a technology should not be a separate entity but integrated into the overall business, product, and operations model. I may be wise at first to establish an “AI center of excellence” in the company. However, it should eventually make itself redundant as AI becomes a natural element of every business function.
The podcast contained many other ideas that I’m covering. I recommend you listen to it yourself! Look for Inference: AI business podcast by Silo AI.
PS. I hope they will start publishing more often. The previous episode on Apple podcasts was from early 2023.
FROM AEC BUSINESS
Recognizing Future AEC Leaders
Awards are an opportunity to pay tribute to the work of future leaders in the AEC sector. Recognizing industry leaders can boost their careers and inspire other young professionals. It also highlights the importance of supporting these champions within companies.
Can Software Vendors Use Your Data for AI Training?
This question underscores a recurring concern within the industry—the need for transparency and clarity in managing user data. As AI and machine learning become increasingly integrated into construction software, understanding the implications of data usage is critical.
领英推荐
SELECTED INDUSTRY NEWS
Consigli Construction leverages Quickbase to solve complex projects
Consigli Construction uses Quickbase to manage complex projects by consolidating data from multiple sources into a single platform. This no-code solution enhances their project management, staffing, planning, and safety procedures. By leveraging Quickbase, Consigli has improved efficiency, streamlined safety documentation, and ensured compliance with rigorous safety standards. The platform allows for real-time access to data, customizable solutions for various job sites, and significant reductions in paperwork and manual updates, helping Consigli maintain high standards and deliver projects on time and within budget.
Work Planning within the Schedule Tool in Autodesk Build
Autodesk Build’s Schedule tool enables teams to gain real-time visibility into project schedules by centralizing schedules in the cloud, allowing teams to collaborate, connect, and interface with the most up-to-date information.
AI-backed robots roll through Seattle construction site, speeding build process and improving safety
AI-backed robots are enhancing safety and efficiency at a Seattle construction site. Skanska, in collaboration with Nextera Robotics, is using autonomous robots to capture high-resolution images, monitor progress, and ensure safety compliance. These robots significantly reduce manual labor, improve project management, and offer real-time insights. This tech marks a pivotal shift in the construction industry, streamlining processes and elevating safety standards. – Source: GeekWire
Architects tap in to Gen AI for design and business efficiencies
Generative AI is transforming architectural practice by enhancing design efficiency and reducing administrative burdens. Tools like Midjourney, Stable Diffusion, and ChatGPT are used for early design visualizations, idea generation, and administrative tasks. These AI applications enable architects to produce visual content rapidly and improve client communication. However, there are concerns about the potential over-reliance on AI and its impact on employment. Custom AI models, tailored for specific tasks and knowledge areas, are being developed to address domain-specific challenges and improve operational efficiency within architectural firms. – Source: The RIBA Journal
Connect Housing Blocks bring modular construction to Columbus housing industry
Connect Housing Blocks is bringing modular construction to Columbus, Ohio, through a new factory to address the housing shortage. The factory will produce up to 1,700 modular housing units annually. The modular units, constructed from light gauge steel, are designed to be fully finished and ready for rapid assembly on-site, significantly reducing construction time and costs. This approach is expected to cut build time by 50% and construction costs by 20%, while also providing superior energy efficiency and reduced waste compared to traditional building methods. Brad DeHays, the founder of Connect Realty, emphasized the regional need for affordable housing and the project’s potential to create 300 new jobs. The project is also notable for its community engagement, partnering with local organizations to recruit and train a diverse workforce, addressing the area’s economic and social needs. – Source: The Columbus Dispatch
New Business Models and Digital Platforms in Construction 4.0
Olivier Lépinoy writes: “All pioneer AEC firms try to reinvent themselves: they collect data, they build their own digital platforms, they operate them with new business models, and they offer new services to their ecosystem. This is an Invisible Revolution because things look unchanged. The most essential changes are changes in the way the industry is structured, not in the way buildings are built.” – Source: Autodesk
UK needs 250,000 extra construction workers by 2028
The Construction Industry Training Board’s annual industry forecast has highlighted the need for extra construction workers to keep up with the rising demand.
AEC BUSINESS PARTNERS
I thank our fantastic partners for supporting AEC Business!
Quickbase
Quickbase is the first application platform built for Dynamic Work Management, empowering nearly 12,000 global organizations to bring together people, processes, and data into one centralized location. Quickbase enhances productivity and reduces Gray Work, the time lost when searching for data and information, by connecting everything through a single source of truth, enabling the right people to have access to the right information, mitigating risk, reducing waste, and trimming unexpected costs. For more information, visit www.quickbase.com .
Solibri
Solibri is the leader in BIM Quality Assurance and Quality Control. Check out Solibri’s AEC Business Channel Impeccable construction .
ProjectReady
ProjectReady is a collaborative project information management solution that brings data and information together to empower professionals in the architecture, engineering, and construction (AEC) space as well as project owners, to make decisions, reduce risk, and save money.
Helping contractors with high-risk contracts | Posts and articles about the process.
5 个月Love this! The insights on how AI is reshaping contech and business models are fascinating.