Home Blog Page 315

Creative ads for life

0

Creativity Beyond the Super Bowl: Why Digital Brands Must Prioritize Originality Year-Round

For anyone in advertising, design, or company building, the Super Bowl has always been a great symbol of creative ambition. Despite ongoing debates about the relevance of traditional advertising, this event remains a cultural touchstone – a rare platform where brands unveil their boldest ideas to captivate millions.

01. Creativity Fuels Adaptability

In today’s digital landscape, where sameness can lead to disengaged audiences and diminishing returns, creative differentiation has become the ultimate competitive advantage. A survey by Deloitte revealed that 87% of businesses identify innovation and creativity as critical to staying competitive and navigating disruption. For digital-first companies, embracing originality means more than just appealing to consumers – it’s about remaining relevant in an ever-evolving marketplace.

02. Creativity Drives Revenue Growth

Investing in creativity isn’t just about standing out – it’s a driver of business success. According to McKinsey & Company, companies identified as "highly creative" outperform their peers in revenue growth by an average of 13% annually. For digital-first brands, the connection should be clear: originality and innovation in product design and user experience not only differentiate a brand in crowded markets but also enable premium pricing.

03. Creativity Builds Brand Loyalty

Creativity isn’t just about catching attention – it’s about forging lasting connections. Brands that deliver distinctive, emotionally engaging experiences are far more likely to earn customer loyalty. According to Adobe’s 2024 State of Creativity report, 64% of consumers believe that a brand’s design reflects how much it values its customers. For digital-first companies, this connection is critical.

Conclusion

By extending the bold creativity seen on Super Bowl Sunday to their digital products and experiences year-round, brands can forge deeper connections with their audiences and drive sustainable growth. In a world craving originality, the only way to stand out – and stay ahead – is to embrace the power of being different.

FAQs

Q: Why is creativity important for digital-first brands?
A: Creativity is essential for digital-first brands to stand out in crowded markets, drive revenue growth, and build brand loyalty.

Q: How does creativity impact business success?
A: Creativity is a driver of business success, with companies identified as "highly creative" outperforming their peers in revenue growth by an average of 13% annually.

Q: What are the benefits of prioritizing creativity?
A: Prioritizing creativity can lead to increased adaptability, revenue growth, and brand loyalty, as well as a competitive advantage in the market.

Blackstone Bullish on A.I. Data Centers

0

Blackstone CEO: AI Systems Don’t Need Huge Amounts of Power and Investment

DeepSeek’s Innovative Approach

The Chinese AI start-up DeepSeek has recently upended the prevailing view that artificial intelligence systems require huge amounts of power and investment. However, Blackstone, the private equity behemoth, remains optimistic.

Blackstone’s President’s Views

"We still think there’s a vital need for physical infrastructure, data centers and power," Jonathan Gray, Blackstone’s president, said during a fourth-quarter earnings call with investors on Thursday. "The form of that use may change."

Adapting to the Changing Landscape

Mr. Gray noted that like much of the investing and corporate world, executives at Blackstone have spent a lot of time in the past week weighing the impact of DeepSeek. The company has made an aggressive push into buying and building data centers, the physical infrastructure used by technology companies to run AI systems. In 2021, it purchased the US data center firm QTS for $10 billion, and last year, it led a roughly $16 billion deal to buy AirTrunk, which operates data centers in Asia.

The Future of AI Adoption

Mr. Gray’s comments echoed the relatively sanguine views of some technology executives in recent days, including those of Microsoft’s chief executive, Satya Nadella. Mr. Gray said he expected that as the cost of computing power for AI drastically decreased, AI would be more widely adopted. In other words, he said, while the amount of power needed for an AI model to answer a specific question may decrease, people will ask more questions.

Blackstone’s Data Center Strategy

Mr. Gray said Blackstone builds data centers only for technology companies that sign long-term leases. "We don’t build them speculatively." The way customers use these data centers, he noted, could very well change.

Blackstone’s Performance

Blackstone’s stock has been on a run lately. It’s traded up roughly 40 percent in the past year, outpacing major stock indexes. It was down roughly 3 percent in midday trading.

A Shift in Focus

Thursday’s investor call was a shift from previous ones. In recent calls, the firm’s senior executives had taken a few minutes at the outset to discuss AI’s importance to the world and Blackstone’s investments around the technology.

Conclusion

Blackstone’s CEO, Stephen A. Schwarzman, has previously described the consequences of AI as "as profound as what occurred in 1880, when Thomas Edison patented the electric lightbulb." He cited expectations that the United States would need approximately $1 trillion of capital expenditures over the next five years for new data centers and another $1 trillion outside the United States.

FAQs

Q: What is DeepSeek’s innovative approach to AI systems?
A: DeepSeek has upended the prevailing view that AI systems require huge amounts of power and investment.

Q: How does Blackstone view the need for physical infrastructure and data centers?
A: Blackstone’s president, Jonathan Gray, believes that there is a vital need for physical infrastructure, data centers, and power, but the form of that use may change.

Q: How does Blackstone’s data center strategy differ from its previous approach?
A: Blackstone builds data centers only for technology companies that sign long-term leases, and the way customers use these data centers could very well change.

Autodesk CTO Raji Arasu calls for diversity in AI teams

0

Raji Arasu: A Pioneer in AI and a Champion for Women in Tech

Raji Arasu, the CTO of Autodesk, has been using AI for “multiple decades” to solve software-related challenges. In an interview with TechCrunch, she shared her insights on the AI revolution and her experiences as a woman in tech.

A Career Spanning Multiple Decades

Arasu has had a long and illustrious career in tech, joining Oracle as an engineering technical lead in 1993. She later worked at eBay, StubHub, and Intuit before joining Autodesk four years ago. Throughout her career, she has been driven by a passion for using AI to solve complex problems.

From Traditional Software Development to AI

Arasu has always turned to AI as a solution when traditional software development approaches have struggled to address complex problems. She has found that AI is particularly effective in handling unstructured data.

Supporting Women in Tech

Arasu believes that a strong professional support network has been essential to her success. She has tried to pay it forward by becoming vocal about supporting women in tech and driving diversity, equity, and inclusion (DEI) initiatives, particularly for boards of directors.

A Front-Row Seat to the AI Revolution

At Autodesk, Arasu has overseen the company’s R&D org, which has explored AI applications in areas like construction automation, 3D modeling, and engineering design. She is excited about the speed at which AI is transforming industries across the board.

Risks and Challenges

Arasu acknowledges that AI comes with risks, including the potential for bias and concerns around data privacy, transparency, and data practices. She believes that close collaboration between industry and government is necessary to create well-informed AI policies.

Building AI Responsibly

Arasu believes that companies must ensure that their AI systems are clear about how they function, so that users can trust or at least not mistrust the decisions they make. This requires understanding customers’ concerns about their data, putting practices in place to protect proprietary and personal information, and maintaining compliance with global policies.

Conclusion

Raji Arasu is a true pioneer in AI and a champion for women in tech. Her experiences and insights offer valuable lessons for anyone interested in the AI revolution and its potential to transform industries across the board.

FAQs

Q: What inspired you to use AI to solve software-related challenges?

A: I’ve always been drawn to AI as a solution when traditional software development approaches have struggled to address complex problems.

Q: How do you think AI can be used to transform industries?

A: AI has the potential to transform industries by automating tasks, improving efficiency, and enabling new business models.

Q: What are some of the risks and challenges associated with AI?

A: AI comes with risks, including the potential for bias, concerns around data privacy, transparency, and data practices.

Q: How can companies build AI responsibly?

A: Companies can build AI responsibly by ensuring that their systems are clear about how they function, understanding customers’ concerns about their data, and maintaining compliance with global policies.

Chameleon LLM: Dynamic Adaptation During Inference

The Problem with Static LLMs

Large language models like GPT-3 and GPT-4 have revolutionized natural language processing (NLP) with their ability to generate human-like text, summarize documents, translate languages, and more. However, these models are typically deployed with fixed weights, meaning they cannot adapt to new or varying data during inference (the phase where the model generates predictions). This static nature can lead to suboptimal performance when the input data differs from what the model was trained on.

What is ChamaleonLLM?

ChamaleonLLM is a framework that enables dynamic adaptation of LLMs during inference. Instead of using fixed weights or pre-learned updates, ChamaleonLLM adapts the model’s behavior on-the-fly based on the statistics of the input batch.

Key Innovations

  • Batch-Aware Clustering: Inputs in a batch are grouped into clusters based on their token embeddings (numerical representations of words or sentences). This clustering ensures that similar inputs are processed together, allowing the model to capture shared context and reduce noise.
  • Dynamic Low-Rank Updates: A hyper-network (a smaller neural network) generates low-rank updates (small adjustments to the model’s weights) tailored to the statistics of each cluster. These updates are computed in real-time, enabling the model to adapt dynamically to the specific characteristics of the input batch.
  • Efficiency: Unlike traditional methods that require storing multiple expert models or masks, ChamaleonLLM generates updates on-the-fly, reducing memory and computational overhead.

How Does ChamaleonLLM Work?

The framework is built on a pre-trained causal language model (e.g., GPT-2) and consists of two main components:

Batch-Aware Clustering

Inputs are tokenized and converted into token embeddings. These embeddings are normalized and grouped into clusters using k-means clustering, a simple algorithm that minimizes the distance between points and cluster centroids. Each mini-batch contains inputs from the same cluster, ensuring that the model processes contextually similar data together.

Adaptive Low-Rank Update Generation

A hyper-network takes the mean token embeddings of each cluster as input and generates low-rank update parameters. These updates are applied to the model’s weights, allowing it to adapt to the specific characteristics of the cluster. The hyper-network is trained to produce updates that improve the model’s performance on the given batch.

Why is ChamaleonLLM Better?

The authors compare ChamaleonLLM with traditional LoRA and unadapted GPT-2 models on the WikiText-2 dataset, a benchmark for language modeling. Here are the key results:

Adaptation Regime Parameters Validation Loss Validation Perplexity
Unadapted GPT-2 124,439,808 10.2513 28,319
Traditional LoRA 204,100 1.3528 3.8683
ChamaleonLLM 6,786,596 0.3753 1.4554

ChamaleonLLM achieves significantly lower validation loss and perplexity compared to traditional LoRA and unadapted GPT-2. The dynamic adaptation mechanism allows the model to generalize better and handle diverse input distributions.

Conclusion

ChamaleonLLM represents a significant step toward making LLMs more flexible and efficient in real-world applications. By enabling dynamic adaptation during inference, this framework can improve the performance of LLMs in scenarios where input data is highly variable or noisy. It also reduces the computational and memory overhead associated with traditional fine-tuning methods.

Frequently Asked Questions

Q: What are the limitations of traditional LLMs?

A: Traditional LLMs are typically deployed with fixed weights, meaning they cannot adapt to new or varying data during inference, which can lead to suboptimal performance.

Q: What is the main innovation of ChamaleonLLM?

A: ChamaleonLLM enables dynamic adaptation of LLMs during inference by leveraging batch-aware clustering and dynamic low-rank updates.

Q: How does ChamaleonLLM improve the performance of LLMs?

A: ChamaleonLLM achieves significantly lower validation loss and perplexity compared to traditional LoRA and unadapted GPT-2 models by enabling dynamic adaptation to the specific characteristics of the input batch.

Q: What are the advantages of ChamaleonLLM?

A: ChamaleonLLM reduces the computational and memory overhead associated with traditional fine-tuning methods and can improve the performance of LLMs in scenarios where input data is highly variable or noisy.

SoftBank’s Son and OpenAI’s Altman Double Down on AI

Masayoshi Son and Sam Altman’s Unprecedented Partnership

When Masayoshi Son and Sam Altman first met in 2017 at SoftBank’s offices in Tokyo, the younger entrepreneur lacked a project large enough to tempt the tech billionaire to open his chequebook. Over the next two years, as Altman’s OpenAI evolved into one of the world’s most potent artificial intelligence companies, Son tried to invest close to $1 billion in the ChatGPT maker. Altman refused.

A $40 Billion Investment

Altman opted for Microsoft, which could provide the computing power needed to turbocharge his company’s growth, according to a person with direct knowledge of the decision. OpenAI has since become one of the fastest-growing start-ups of all time and is straining to grow beyond its partnership with Microsoft. This has cleared a path for SoftBank to lead the largest-ever start-up investment – a $40 billion round into OpenAI. In return, Son gets something he craves: a place at the center of the AI world.

Stargate: A $500 Billion AI Data Center Project

On stage in Tokyo, the two men discussed a connected plan: a $500 billion AI data center project dubbed "Stargate." Son, a few years shy of 70, still reveling in his abilities as a showman, sat across from Altman, nearly 30 years his junior and a man who seemed less at ease in front of a crowd.

The Conviction of AGI

For all their differences, they share a conviction that the era of artificial general intelligence – when machines can outperform humans across cognitive fields – is at hand, and that exponentially more powerful superintelligence will follow.

The Deal

The deal Son and Altman are putting together is unprecedented in size and structure, with talks between SoftBank and OpenAI playing out over two stages, according to people with knowledge of the plans. First, there is an exclusive funding round in which SoftBank has committed to invest up to $40 billion in OpenAI. SoftBank is not expected to provide the full amount on its own and has the right to bring other partners into the funding round. OpenAI can approve or veto the new investors, according to a person with direct knowledge of the plans.

Stargate: A Joint Venture

In the second phase, SoftBank is planning to initially put about $15 billion to $20 billion into Stargate. OpenAI, with new money from the SoftBank round, has agreed to match whatever figure SoftBank puts into the infrastructure project. Partners including MGX, the Abu Dhabi state-owned fund, and a Stargate equity investor, intend to make further investment decisions once they see progress on the project, according to three people with knowledge of the plans.

Conclusion

The deal fits Son’s record of taking leaps of faith, backed by billions in capital, often overriding the recommendations of his lieutenants. His search for the next big thing has also resulted in numerous dramatic losses, including $14.2 billion on Adam Neumann, the founder of WeWork, which eventually tumbled into bankruptcy.

FAQs

Q: What is Stargate?
A: Stargate is a $500 billion AI data center project.

Q: How much is SoftBank investing in Stargate?
A: SoftBank is planning to initially put about $15 billion to $20 billion into Stargate.

Q: What is OpenAI’s valuation after the SoftBank investment?
A: OpenAI’s valuation is expected to be around $260 billion after the SoftBank investment.

I Don’t Want iPhone SE 4

0

The iPhone SE: A Nostalgic Outlier in Apple’s Lineup

The iPhone SE has always been something of an outlier in the Apple lineup. Sticking rigidly to an older design, the budget phone carries an air of nostalgia, particularly for a time when phones didn’t require at least three hands to hold.

Rumors of the SE 4

If all of the design rumors about the SE 4, reportedly due to be announced next week, are true, I’ll be bitterly disappointed. The days of the ‘small phone’ have been spluttering to an end for a while now. After reportedly low sales, Apple unceremoniously ditched the ‘iPhone mini’ lineup after just two years in 2022, replacing it with just the opposite – the enormous ‘iPhone Plus’. In that landscape, the only option for those who prefer something truly pocketable was the iPhone SE 3, with its iPhone 8-inspired design. But the reported adoption of an iPhone 14-inspired body for this month’s SE 4 suggests Apple will be slamming the door on several features; the Home Button, Lightning port and, most devastatingly for the short-fingered among us, a small size.

A Fan-Made Render of the iPhone SE 4

A fan-made render of the iPhone SE 4
(Image credit: AppleTrack )

The End of Small Phones?

If the iPhone SE does adopt a 6.1-inch display, and every rumor suggests it will, then the era of small phones will well and truly be over. The budget phone is rumored to feature a single camera lens, 60hz display and USB-C port, bringing several contemporary features to the SE model (although it reportedly won’t get a Dynamic Island, just a notch).

A Hand Holding an iPhone 13

A hand holding an iPhone 13

I miss the iPhone 13 mini
(Image credit: Jeremy Bezanger via Unsplash)

A Perfect World

But while the need to refresh the internals of the SE make sense, particularly with the advent of Apple Intelligence, I still don’t want the SE 4 to happen. Or at least this version of it. In a perfect world (by which I mean my own very specific perfect world), Apple would use the body of the iPhone 13 mini for the SE 4. It’s no secret that phones have become increasingly addictive, and in my own deeply unscientific opinion, those immersive screens probably aren’t helping. The mini was perfect for those of us who see the phone as a tool rather than a portable IMAX display; something to be whipped out of a pocket when needed, then stowed away and forgotten about. Something that isn’t a bright, shiny brick.

The iPhone SE 4: A Rumored Design

iPhone 14

The iPhone SE 4 is strongly rumoured to resemble the iPhone 14
(Image credit: Apple)

Conclusion

Alas, it seems I’m almost alone in my views. The sales figures for the iPhone 13 mini speak for themselves – that is, they were apparently dire. The people simply don’t want small phones. So before the era of small devices finally ends, you might want to snap up the iPhone SE 3 while stocks last. Because in the words of D:Ream, things can only get bigger.

FAQs

Q: What are the rumored features of the iPhone SE 4?
A: The iPhone SE 4 is rumored to feature a 6.1-inch display, single camera lens, 60hz display, and USB-C port.

Q: Will the iPhone SE 4 have a Dynamic Island?
A: No, the iPhone SE 4 is rumored to have a notch instead of a Dynamic Island.

Q: Will the iPhone SE 4 be available in a small size?
A: No, the iPhone SE 4 is rumored to have a larger display, similar to the iPhone 14.

Q: Why do you think the iPhone SE 4 should have a smaller size?
A: I believe that smaller phones are more practical and easier to use, and that the iPhone SE 4 should cater to those who prefer a more compact device.

Treasury Records Blocking

0

Temporary Block on DOGE Access to Treasury Department Records

A federal judge has temporarily blocked DOGE, a company founded by Elon Musk, from accessing Treasury Department records containing personal information such as Social Security numbers. The order was issued in response to a lawsuit filed yesterday in New York’s Southern District Court.

Background of the Lawsuit

The lawsuit alleges that the administration exceeded its authority, broke the US Administrative Procedures Act, and violated the US Constitution’s separation of powers doctrine when it granted DOGE access to the Treasury’s federal payments system. The suit was filed by New York Attorney General Letitia James and joined by attorneys general from 18 other states.

Order from District Judge Paul A. Engelmeyer

District Judge Paul A. Engelmeyer’s order prohibits the Trump administration from granting access to Treasury Department payment records, payment systems, or other data systems containing personally identifiable information and/or confidential financial information of payees to political appointees, special government employees, and government employees detailed from an agency outside the Treasury Department.

The order also requires those prohibited from accessing the records to immediately destroy any and all copies of material downloaded from the Treasury Department’s records and systems since January 20th, 2025. A hearing is scheduled for February 14th.

Reactions to the Order

New York Attorney General Letitia James stated in a press release that President Trump does not have the power to give away Americans’ private information to anyone he chooses, and he cannot cut federal payments approved by Congress. Musk and DOGE have no authority to access Americans’ private information and some of the country’s most sensitive data.

Conclusion

The temporary block on DOGE’s access to Treasury Department records is a significant development in the ongoing legal battles between the administration and state attorneys general. The order highlights the importance of protecting Americans’ private information and upholding the rule of law.

FAQs

Q: What is the lawsuit about?
A: The lawsuit alleges that the administration exceeded its authority, broke the US Administrative Procedures Act, and violated the US Constitution’s separation of powers doctrine when it granted DOGE access to the Treasury’s federal payments system.

Q: Who filed the lawsuit?
A: New York Attorney General Letitia James and attorneys general from 18 other states filed the lawsuit.

Q: What does the order prohibit?
A: The order prohibits the Trump administration from granting access to Treasury Department payment records, payment systems, or other data systems containing personally identifiable information and/or confidential financial information of payees to political appointees, special government employees, and government employees detailed from an agency outside the Treasury Department.

Q: What is the next step?
A: A hearing is scheduled for February 14th to further discuss the lawsuit and the order.

Adobe Acrobat’s AI Assistant Deciphers Complex Contracts

Understanding Contracts Just Got Easier with Adobe’s New AI-Powered Tool

Introducing the Future of Contract Analysis

All of us have to deal with contracts and agreements as consumers or business professionals. However, as most contracts are written in legalese, understanding them can be challenging. To ease the pain, Adobe has added a new skill to its Acrobat AI that tries to interpret complex contracts for the average person.

How it Works

Automatically detecting when a PDF is a contract, the AI can summarize the information, suggest questions for you to ask, and even find differences between multiple versions of a contract. To access the new contract-reading superpower, you’ll need to purchase Adobe’s AI Assistant for Acrobat, a $5 per month add-on for the paid Adobe Acrobat and the free Acrobat Reader.

Key Features

  • Recognizes regular PDFs and scanned documents set up as contracts
  • Generates an overview and highlights keywords and terms for you
  • Provides a brief summary of the details with cleaner, clearer language and citations
  • Scans for differences among multiple versions of the same contract and finds inconsistencies and discrepancies
  • Allows you to review the contract with other people and parties involved and request e-signatures directly from Adobe Acrobat or Reader

Using the New Skill

You can use the new skill for a variety of contract types, including credit card and vendor agreements, loyalty programs, purchase orders, and business agreements. For example, a business owner could use the AI to review a partnership agreement with their attorney. A financial analyst can review sales contracts. Consumers could use it to analyze changes in everything from apartment leases to mobile phone plans.

Potential Pitfalls

AI is far from perfect. How do you know that Adobe’s AI is analyzing and interpreting the contract correctly? To address this concern, Adobe says that its AI assistant combines its large language models (LLMs) with the same artificial intelligence and machine learning models behind its Liquid Mode technology.

Data Security

AI is also infamous for training itself on user data. Will your contracts be on the training list? To address this concern, Adobe says that the AI is governed by data security protocols and developed in line with its AI ethics processes. The company promises that it will never train its generative AI models on customer data and restricts third-party LLMs from similar training.

Conclusion

Adobe’s new AI-powered contract analysis tool is a game-changer in helping simplify the experience of understanding complex contracts. With its ability to summarize information, suggest questions, and find differences between multiple versions, it can make it easier for customers to navigate the often-complex world of contracts.

FAQs

Q: What is the cost of the AI Assistant for Acrobat?
A: The AI Assistant for Acrobat is a $5 per month add-on for the paid Adobe Acrobat and the free Acrobat Reader.

Q: What types of contracts can I use the new skill for?
A: You can use the new skill for a variety of contract types, including credit card and vendor agreements, loyalty programs, purchase orders, and business agreements.

Q: Is the AI perfect?
A: No, AI is far from perfect. However, Adobe’s AI assistant combines its large language models with the same artificial intelligence and machine learning models behind its Liquid Mode technology to offer a more accurate analysis of the document’s structure and content.

Q: Will my contracts be used to train the AI?
A: No, Adobe promises that it will never train its generative AI models on customer data and restricts third-party LLMs from similar training.

Engineering joy | MIT News

0

When the late professor emeritus Woodie Flowers SM ’68, MEng ’71, PhD ’73 was a student at MIT, most of his classes involved paper-and-pencil exercises with predetermined solutions. Flowers had an affinity for making things, and for making them work. When he transitioned from student to teacher, he chose to carry this approach into his method of instruction and, in doing so, he helped change the way engineering students are educated — at MIT, and around the world. 

Flowers passed away in 2019, but his legacy lives on, and the magnitude of the educational revolution he helped to evolve was profound.

In the 1970s, Flowers took over instruction of 2.70, now called class 2.007 (Design and Manufacturing I). The capstone course is one that many first-year students today look forward to taking, but that wasn’t always the case. Before Flowers took over instruction, class instruction relied heavily on chalkboard demonstrations.

“Their idea of design at the time was to draw drawings of parts,” explains Professor Emeritus David Gossard PhD ’75, Flowers’ longtime friend and colleague. “Woody had a different idea. Give the entire class a kit of materials [and] a common goal, which was to build a machine — to climb a hill, or pick up golf balls, or whatever it did — and make a contest out of it. It was a phenomenal success. The kids loved it, the faculty loved it, the Institute loved it. And over a period of years, it became, I think it’s fair to say, an institution.”

With Flowers at the lead, 2.70 transformed into a project-based, get-your-hands-dirty, robotics-competition-focused experience. By all accounts, he also made the experience incredibly fun — something he valued in his own life. He was fond of skydiving and was often seen rollerblading through the Infinite Corridor. The course, informed by his unique style, was at the forefront of a revolution in engineering education, and it quickly helped solidify the Department of Mechanical Engineering’s reputation for innovative education.

“A lot of kids had never started from scratch and built anything,” Flowers once told The Boston Globe. His advisor, Robert Mann, had similar beliefs in a hands-on, modern pedagogy. Building on Mann’s philosophy, and incorporating his own approach, Flowers breathed new life and provided a new foundation for “the MIT way” of teaching. This was a reinvigoration at the right place and the right time that ultimately had a global butterfly effect on the popularity of science, technology, engineering, and math (STEM) instruction.

“Over the years lectures had displaced the hands-on stuff, and Woodie brought it back,” says Sanjay Sarma, the Fred Fort Flowers (1941) and Daniel Fort Flowers (1941) Professor in Mechanical Engineering. “I can’t think of a single person to have impacted the field of robotics and design in undergraduate, or high school, education as much as Woodie.”

Flowers became interested in mechanical engineering and design at a young age, thanks in large part to his parents. His father was a welder with a penchant for tinkering, inventing, and building, his mother was an elementary school teacher. Flowers grew up taking things apart and putting them back together — an activity which he seemed to believe made students better engineers.

Speaking in 2010 with InfiniteMIT, a digital archive of Institute history made possible by the generosity of Jane and A. Neil Pappalardo ’64, Flowers shared a story about a student who had accepted the task in her group of finding out whether a piece of reinforcement steel rebar could be bent into a tight loop and serve as a bearing.

“She came into lab and I was there early, and she had a slightly bent piece of rebar. It had been heated — you could tell that it had been hot, and she was going to report that she really can’t do that, it just kind of doesn’t work,” Flowers recalled. He suggested they try another approach.

“We went out in the lab and I found another big steel bar and I found the biggest vice I could find,” he continued. Flowers cranked the rebar down against the piece of steel he was going to wrap it around, then took a four-pound sledgehammer to it. “My father had a blacksmith pit, so that was familiar to me. I wrapped [the rebar around the steel and] made a fine bearing. As I finish the last blow, I looked up and three of the best students in the class — really sharp people — were standing there with their jaw open. They’d never seen anyone hit a piece of steel hard enough to just mold it.”

He continued, “that visceral understanding of the behavior of mechanics is really important. It doesn’t fall out of the sky and it certainly doesn’t come out of a textbook, it comes through real interaction. I believe I had been so lucky because when I encountered Castiglione’s theorem about deflection of materials, it kind of made sense.”

Course 2.70/2.007 is considered a landmark class in engineering education. It was one of the first hands-on classes to teach students not only how to design an object, but also how to build it and, by demonstrating the value of practical, project-based learning and robotics competitions, it has influenced the approach taken by many other programs. Today, it continues to develop students’ competence and self-confidence as design engineers, with an emphasis on the creative design process bolstered by application of physical laws, robustness, and manufacturability.

Notably, the course also served as the inspiration for development of the FIRST Robotics program, which Flowers and inventor Dean Kamen started in 1989. FIRST has programs for preschool through high school students and, to date, more than 3.2 million youth from more than 100 countries have participated in FIRST competitions.

In the 1970s, the parts kit — or as Flowers fondly referred to it, the “bag of junk” — included things like springs, tongue depressors, and rubber bands. Flowers’ wife Margaret recalls spending many nights packing these kits and hosting advisees in their home. “We considered ourselves a team,” she says.

Today, in addition to using the kit of mechanical parts and materials, students in 2.007 might develop 3D printed components, and they incorporate electronics in their robots for an autonomous portion of the final competition.

The spring 2024 competition, themed after Cartoon Network’s popular animated science fiction sitcom “Rick and Morty,” featured a spaceship that students’ robots could interact with for points, vats of “acid” where balls could be collected and placed in tubes, and game pieces that paid homage to iconic episodes. The final task required the robot to travel up an elevator and send a character down a zipline.

In recent years, other themes have centered on tasks related to stories ranging from “Star Wars” to “Back to the Future” and “Wakanda Forever.” The 2022 theme, however, may have been the most poignant theme to date: “Legacy,” a celebration of Flowers’ life and work.

“[Woodie] revealed, unambiguously, that designing, fabricating, assembling and building things was fun,” says Gossard. “It was arguably the essence of engineering. There was joy in it.”

A version of this article appears in the Spring 2025 issue of MechE Connects, the magazine of the MIT Department of Mechanical Engineering.