We don’t often cover telecom technology startups, but it’s periodically worth checking in to see what’s happening in that space. We can get a good indication from the latest cohort to emerge from an accelerator associated with South Korea’s largest wireless carrier, SK Telecom.
This group of startups will join the Telecom Infra Project accelerator in South Korea, which is part of a global program of telecoms specialist centers, and run in partnership with SK Telecom.
The cohort includes a ship berthing monitoring system; an app that turns a group of mobile phones into a TV studio; an AI-powered indoor positioning system which creates interactive maps; a vision system for delivery robots; and one which allows remote audiences to experience live events ‘together’ via a digital stadium.
The selected start-ups include:
Seadronix: Seadronix is a computer vision-based ship parking monitoring solution that provides an AI-based berthing monitoring system. Through producing around-view images, distance and speed between the ship and the port to pilot or captain it assists ship parking. Crunchbase
39degC: This is a mobile multi-camera live streaming app. It directly connects multiple smartphone feeds to each other using a technology called WiFi-Direct – turning them into a TV studio.
Kiswe: Kiswe is a supplier of entertainment broadcast technology. Its product CloudCast is a “Broadcast Studio in the Cloud” which enables partners to send a digital feed into the cloud to produce live and non-live content. Its other product, Hangtime , allows remote audiences to experience live events ‘together’ through creating a digital stadium with chat rooms, and control over viewing angles from within the platform. Crunchbase
Support Indie Bookstores With Your Holiday Shopping
The pandemic hasn’t been kind to independent bookstores. While indies have handled competition from Amazon and big-box sellers relatively well in recent years, having to shut their doors due to COVID-19 and respond to customers’ desire for online ordering has not been easy—and despite the American Association of Publishing noting steady book sales amid the pandemic, bookstore sales dropped 30.7 percent in August, according to industry newsletter Shelf Awareness. If you want to keep indie bookstores alive, consider directing some of your holiday gift-buying budget toward your local independent bookseller. We’ve rounded up a number of ways you can support indies this season.
Heads up, though: if you are going to gift books and book-related items, start shopping and ordering now. Indie booksellers can always use your support—and ordering earlier can help them manage their inventory and meet demand ahead of the holidays.
Order directly from your local indie
Some booksellers responded to the pandemic by setting up shop online and offering pickup, local delivery, or both. Search for indie bookshops in your neighborhood or city to find out what they’re doing. IndieCommerce has a directory of independent sellers that handle e-commerce.
Shop on IndieBound or Bookshop
If you can’t easily get a book from a store up the street, try IndieBound or Bookshop instead of defaulting to Amazon. IndieBound has a zip code search that directs you to local booksellers that have your desired title(s) in stock, while Bookshop allows you to either select a local seller to receive 100% of the profit from your order or to put your money into a pool that’s distributed among member shops.
Note: IndieBound will also connect you to Bookshop if it can’t find titles in stock in your area. Twitter users have suggested that Bookshop’s shipping times are actually comparable to Amazon’s and that the customer service is (unsurprisingly) far better.
Shop for non-book gifts
Many, if not most, indie bookstores sell more than just books. As writer Celeste Ng points out, you can order jigsaw puzzles, apparel, tote bags, and other book-adjacent gifts (coffee mugs with a literary quote, anyone?). Try to find all of your stocking stuffers–bookmarks, pens, magnets, socks—there too, along with your holiday cards.
Gift a subscription to Libro.fm
If audiobooks are more your friends’ and family’s thing, gift them a membership to Libro.fm. The platform allows you to choose a local, independent bookstore to receive a cut of the profit from each subscription. You can also gift individual audiobooks.
Gift an indie seller gift card
If you’re not sure about a gift recipient’s taste in books, send them a gift card to their local indie or your local indie.
Put together DIY book packages
Get creative with your gift-giving with themed book and book-adjacent packages. This Twitter thread from author Rebecca Makkai has a long list of ideas, from a city-specific gift box to a food-and-book (wine-and-book?) pairing, to a book and film package from the recipient’s birth year. A few more we like:
- Long-distance book club: Send the same book to a group of friends along with an invite to a Zoom gathering and a cocktail recipe card.
- Cookbook and spice package: Gift your food-loving friends a cookbook plus any hard-to-find ingredients.
- Free Little Library: Gift a Free Little Library box filled with wrapped books to a friend who has a yard.
Many indies will wrap orders, too, so you can have items sent directly to friends and family.
Stop by for a cup of coffee
If your indie bookseller is open and serving coffee and food, stop in when you can (and follow all safety rules when you do).
Learn Python machine learning with these essential books and online courses
Teaching yourself Python machine learning can be a daunting task if you don’t know where to start. Fortunately, there are plenty of good introductory books and online courses that teach you the basics.
It is the advanced books, however, that teach you the skills you need to decide which algorithm better solves a problem and which direction to take when tuning hyperparameters.
A while ago, I was introduced to Machine Learning Algorithms, Second Edition by Giuseppe Bonaccorso, a book that almost falls into the latter category.
While the title sounds like another introductory book on machine learning algorithms, the content is anything but. Machine Learning Algorithms goes to places that beginner guides don’t take you, and if you have the math and programming skills, it can be a great guide to deepen your knowledge of machine learning with Python.
Oiling your machine learning engine
Machine Learning Algorithm kicks off with a quick tour of the fundamentals. I really liked the accessible definitions Bonaccorso uses to explain key concepts such as supervised, unsupervised, and semi-supervised learning and reinforcement learning.
Bonaccorso also draws great analogies between machine learning and descriptive, predictive, and prescriptive analytics. The machine learning overview also contains some hidden gems, including an introduction to computational neuroscience and some very good precautions on the pitfalls of big data and machine learning.
That said, the machine learning overview does not go into too much details and would be hard to understand for novices. Given the audience of the book, it serves to refresh and solidify your understanding of machine learning, not to teach you the basics.
Next, Machine Learning Algorithms builds up on that brief overview and goes into more advanced concepts, such as loss functions, data generation processes, independent and identically distributed variables, underfitting and overfitting, different classification strategies (one-vs-one and one-vs-all), and elements of information theory. Again, the definitions are smooth and very accessible for someone who has already had hands-on experience with machine learning algorithms and linear algebra.
Before going into the exploration of different algorithms, the book covers some more key concepts such as feature engineering and data preparation. Here, you’ll get to revisit some of the key classes and functions of scikit-learn, the main Python machine learning library. If you already have a solid knowledge of Python and numpy, you’ll find this part a pleasant review of one-hot encoding, train-test splitting, imputing, normalization, and more. There is some very great stuff in the third chapter, including one of the best and most accessible definitions of principle component analysis (PCA) and feature dependence in machine learning algorithms. You’ll also get to see some of the more advanced techniques not covered in introductory books, such as non-negative matrix factorization (NNMF) and SparsePCA. Of course, without the background in Python machine learning, these additions will be of little use to you.
The real meat ofthe book starts in the fourth chapter, where you get to the machine learning algorithms. Here, I had mixed feelings.
A rich roster of machine learning algorithms
In general, Machine Learning Algorithms is nicely structured and stands up to the name. There are chapters on regression, classification, support vector machines (SVM), decision trees, and clustering. The book follows up with a few chapters on recommendation systems and natural language processing applications, and finishes off with a very brief overview of deep learning and artificial neural networks.
The main chapters offer in-depth coverage of principle machine learning algorithms in Python, including details not covered in introductory books. For instance, the regression chapter goes into an extensive coverage of outliers and methods to mitigate their effects. The classification chapter has a nice discussion on passive-aggressive classification and regression in online algorithms. The SVM chapter has a comprehensive (but complicated) discussion on semi-supervised vector machines. And the decision trees chapter provides a good coverage of the specific sensitivities of DTs such as class imbalance, and some practical tips on tweaking trees for maximum performance.
The clustering section really shines. It spans across three full chapters, starting with fundamentals (k-nearest neighbors and k-means) and goes through more advanced clustering (DBSCAN, BIRCH, and bi-clustering) and visualization techniques (dendrograms). You’ll also get a full account of measuring the effectiveness of the results and determining whether your algorithm has latched onto the right number and distribution of clusters.
Across the book, there are thorough discussions of the mathematical formulas behind each machine learning algorithm. You need to come strapped with solid linear algebra and differential and integral calculus fundamentals to fully understand this (if you need to hone your machine learning math skills, I’ve offered some guidance in a previous post).
The book also makes extensive use of functions numpy, scipy, and matplotlib libraries without explaining them, so you’ll need to know those too (you can find some good sources on those libraries here).
One of the most enjoyable things about Machine Learning Algorithms are the chapter summaries. After going through the nitty-gritty of the math and Python coding of each machine learning algorithm, Bonaccorso gives a brief review of where to apply each of the techniques presented in the book. There are also many references to relevant papers that provide more in-depth coverage of the topics discussed in the book. It’s refreshing to see some of the old but fundamental papers from early 2000s being mentioned in the book. Those things tend to get buried under the hype surrounding state-of-the-art research.
Machine Learning Algorithms finishes off with a good wrap-up of the machine learning pipeline and some key tips on choosing between the different Python tools introduced across the book.
Not enough real-world examples
The one thing, in my opinion, that should set a book on Python machine learning apart from research papers and theoretical textbooks are the examples. A good book should be rich in use-case oriented examples that take you through real-world applications and possibly build up through the book.
Unfortunately, in this respect, Machine Learning Algorithms leaves a bit to desire.
For one thing, the examples in the book are mostly generic, using data-generation functions in scikit-learn such as make_blobs, make_circles, and make_classification. Those are good functions to show certain aspects of Python machine learning, but not enough to give you an idea of how to use the techniques in real life, where you have to deal with noise, outliers, bad data, and features that need to be normalized and categorized.
The code is in plain Python scripts as opposed to the preferred Jupyter Notebook format (which is not much of a big deal, to be fair). Also, while the book omits much of the sample code and focuses on the important parts for the sake of brevity, it made it hard to navigate the sample files at times.
The book does cover some real-world examples, including one with airfoil data in the SVM chapter and another with the Reuters corpus in the NLP chapter. The recommendation systems chapter also includes a few decent use cases, but that’s about it. Without concrete examples, the book often reads like a disparate reference manual with code snippets, which makes it even more crucial to have solid experience with Python machine learning before picking this one up.
Another thing that didn’t really appeal to me were the two chapters on deep learning. Machine Learning Algorithms provides a good overview of deep learning and discusses convolutional neural networks, recurrent neural networks, and other key architectures. But the problem is that introductory books on Python machine learning already cover these concepts and much more. So most of the people who make it this far through the book without putting it down won’t find anything new here (aside from the mention of KerasClassifier maybe).
Midway through Python machine learning journey
So, where does this book stand in the roadmap to learning machine learning with Python? It’s neither beginner level, nor super-advanced. I would suggest picking up Machine Learning Algorithms after you read an introductory-to-intermediate book like Python Machine Learning or Hands-on Machine Learning, or an online course like Udemy’s “Machine Learning A-Z.” Otherwise, you won’t be able to make the best of the rich content it has to offer.
Once you finish this one, you might want to consider Bonaccorso’s Mastering Machine Learning Algorithms, Second Edition, which expands on many of the topics presented in this book and takes them into even greater depth.
This article was originally published by Ben Dickson on TechTalks, a publication that examines trends in technology, how they affect the way we live and do business, and the problems they solve. But we also discuss the evil side of technology, the darker implications of new tech and what we need to look out for. You can read the original article here.
Inside Foxconn’s empty buildings, empty factories, and empty promises in Wisconsin
Hopes were high among the employees who joined Foxconn’s Wisconsin project in the summer of 2018. In June, President Donald Trump had broken ground on an LCD factory he called “the eighth wonder of the world.” The scale of the promise was indeed enormous: a $10 billion investment from the Taiwanese electronics giant, a 20 million-square-foot manufacturing complex, and, most importantly, 13,000 jobs.
Which is why new recruits arriving at the 1960s office building Foxconn had purchased in downtown Milwaukee were surprised to discover they had to provide their own office supplies. “One of the largest companies in the world, and you have to bring your own pencil,” an employee recalls wondering. Maybe Foxconn was just moving too fast to be bothered with such details, they thought, as they brought their laptops from home and scavenged pencils left behind by the building’s previous tenants. They listened to the cries of co-workers trapped in the elevators that often broke, and noted the water that occasionally leaked from the ceiling, and wondered when the building would be transformed into the gleaming North American headquarters an executive had promised.
The renovations never arrived. Neither did the factory, the tech campus, nor the thousands of jobs. Interviews with 19 employees and dozens of others involved with the project, as well as thousands of pages of public documents, reveal a project that has defaulted on almost every promise. The building Foxconn calls an LCD factory — about 1/20th the size of the original plan — is little more than an empty shell. In September, Foxconn received a permit to change its intended use from manufacturing to storage.
Even the handful of jobs the company claims to have created are less than real: many of them held by people with nothing to do, hired so the company could reach the number required for it to get tax subsidy payments from Wisconsin. Foxconn failed at that objective, too: last week, Wisconsin rejected the company’s subsidy application and found it had employed only 281 people eligible under the contract at the end of 2019. Many have since been laid off.
Foxconn did not return repeated requests for comment.
It’s not unusual for either the Trump administration or Foxconn to make announcements that prove hollow. But for Foxconn, the show went on — for two years, the company, aided by the vocal support of the Wisconsin GOP, worked to maintain an illusion of progress in front of a business venture that never made economic sense.
That illusion has had real costs. State and local governments spent at least $400 million, largely on land and infrastructure Foxconn will likely never need. Residents were pushed from their homes under threat of eminent domain and dozens of houses bulldozed to clear property Foxconn doesn’t know what to do with. And a recurring cycle of new recruits joined the project, eager to help it succeed, only to become trapped in a mirage.
Months after the 2018 groundbreaking, the company was racing to hire the 260 people needed to receive the first tranche of payments from the lucrative subsidy package passed by then-Gov. Scott Walker. Recruiters were told to hit the number but given little in the way of job descriptions. Soon, the office began to fill with people who had nothing to do. Many just sat in their cubicles watching Netflix and playing games on their phones. The reality of their situation became impossible to ignore. Multiple employees recall seeing people cry in the office. “The best is when you’re in the elevator with somebody and then they just scream out of nowhere,” said an employee who experienced this several times. “They’ve had enough, because things don’t make sense here.”
“Imagine being in a job where you don’t really know if it’s real or not. Or you know it’s not real, but you don’t know it’s not real. It’s a constant thing you’re doing in your head day after day,” said one employee, who returned to the rented building Trump had spoken at, where workers had been assembling TVs, only to find the line shut down and the lights dimmed a couple of weeks after the photo op was over. “I think all of us were on the verge of a major breakdown.”
It was just the beginning. Foxconn would spend the next two years jumping from idea to idea — fish farms, exporting ice cream, storing boats — in an increasingly surreal search for some way to generate money from a doomed project. Frequent leadership changes, a reluctance to spend money, and a domineering corporate culture would create an atmosphere employees described as toxic. Many of the employees The Verge spoke with have since left the company, and all of them requested anonymity out of fear of retaliation. It has been a baffling ordeal for the people who thought they were building the Silicon Valley of the Midwest — “Wisconn Valley,” Walker called it — all the more so because so many others still believe the vision.
“All people see is the eighth wonder of the world,” said an employee. “I was there and it’s not real. I mean, it’s not. This is something I can’t talk about ever again, because people think you’re crazy, like none of this could ever happen. How could this happen in the US?”
Foxconn’s Wisconsin saga began two days after Trump’s inauguration, when the company’s founder and CEO, Terry Gou, told reporters he was considering building a $7 billion factory in the US and employing as many as 50,000 people.
Such announcements are far from unusual for Gou, and often, nothing comes of them. In Vietnam in 2007, in Brazil in 2011, in Pennsylvania in 2013, and in Indonesia in 2014, Foxconn announced enormous factories that either fell far short of promises or never appeared. Just this year, the industries minister of Maharashtra, India, which aggressively pursued one of Gou’s multibillion-dollar projects in 2015, finally confirmed the factory isn’t coming, saying the state had learned a lesson about believing businesses promising big investments.
In China, where Foxconn employs the vast majority of its million workers, these sorts of announcements are called “state visit projects,” according to Willy Shih, a Harvard business school professor and former display industry consultant. Officials get a ribbon-cutting photo op, the company gets political goodwill, and everyone understands that the details of the contract are just an opening bid by a company that will ultimately do whatever makes economic sense. And in 2017, Gou, like many manufacturing companies with a dependency on China, had an obvious need for goodwill from Trump, who was threatening a trade war that Gou later told shareholders was Foxconn’s “biggest challenge.”
From a political perspective, Gou could hardly have done better than the Wisconsin deal: a factory in a swing state Trump narrowly won, where the Republican governor was campaigning for reelection on job creation, in the district of then-Speaker of the House Paul Ryan and Wisconsin State Assembly Speaker Robin Vos. From a business perspective, the project made less sense. Experts noted an impending glut in LCDs that would drive down prices, high local labor costs that would obliterate the industry’s thin margins, and a lack of key suppliers like the glass-maker Corning that would have to be paid to come to Wisconsin as well. Undeterred, the Trump administration arranged an April meeting between Gou and Walker, who proceeded to pursue Foxconn aggressively.
Even then, Foxconn was highly fluid about what exactly it was going to build. The company had originally floated two possible factories, a “generation 10.5 fabrication facility,” or “Gen 10.5 Fab,” which would manufacture large LCD screens and employ 8,800 people, or a smaller Gen 6 to make smaller LCDs and employ 5,200 people, according to a request for proposal obtained by The Verge. But Foxconn wanted $3 billion in subsidies, according to a source involved with the project, and Walker wanted a five-figure jobs number.
As Gou and Walker met over an early July visit to Wisconsin, the plan grew to nearly the size of both proposed factories combined, employing 13,000 people, and the state’s subsidy swelled from $1.5 billion to $3 billion, according to WEDC correspondence obtained in a records request. No one, according to the source, examined whether what Foxconn was proposing was commercially viable. “There was this assumption that they’re one of the biggest companies out there,” the source said. “Surely they know what they’re doing.” Those were the numbers written onto the single sheet of stationery Walker signed on July 12th to kick off the deal.
A contract with WEDC signed in November made it official: nearly $3 billion in “refundable” tax credits, most likely to be made in the form of direct payments to Foxconn. Combined with infrastructure the state promised to build, approximately $800 million in additional incentives mostly from the small town of Mount Pleasant, where the “Fab” was to be located, and other contributions, the package totaled more than $4 billion. In a best-case scenario, the Legislative Fiscal Bureau found the state wouldn’t break even until 2043. Depending on how many people Foxconn hired, each job would cost taxpayers somewhere between $200,000 and more than a million dollars. The average subsidy in the US is around $24,000 per job.
By the time Trump, Ryan, Walker, and Gou convened for the groundbreaking ceremony in Mount Pleasant on June 28th, 2018, Foxconn had already begun to waver. Days before, the company admitted that rather than building the Gen 10.5 Fab specified in its contract, it would instead build the smaller Gen 6, but still create 13,000 jobs. Two months later Foxconn executive Louis Woo told the local press there was another change: rather than manufacturing jobs, 90 percent of the roles would be researchers and engineers, a number that would mean Foxconn was creating a research campus in Southeastern Wisconsin three times the size of MIT. Woo said these researchers would be developing Foxconn’s “AI 8K+5G ecosystem,” something that, other than being a list of different technologies, has never been coherently explained.
Though Foxconn’s plans appeared in flux, state and local governments continued work on the infrastructure Foxconn had said it would need. Mount Pleasant’s village board voted to blight the area slated for the factory and threatened residents with eminent domain if they didn’t move.
By the end of the summer, Walker found himself in a tight reelection race against state school superintendent Tony Evers, a critic of the deal. Polling showed that few people felt the project would benefit their local economy, so Walker campaigned to show that all of Wisconsin would feel the effect of the “Foxconn bonus.” He was aided in this message by a string of announcements Foxconn: a promised gift of $100 million to the University of Wisconsin-Madison; partnerships with local companies; and the purchase of buildings in far corners of the state that would become “innovation centers,” which Walker quickly featured in campaign ads. In September, Foxconn released a video rendering of its plans for Mount Pleasant: a futuristic corporate campus where employees read books as they took self-driving cars to work in glass orbs.
It wasn’t enough. Walker lost.
The election results fell like “a dark cloud” over Foxconn’s Milwaukee headquarters, in the words of one employee. The company had little progress to show the new administration. Rather than the 1,040 people Foxconn intended to hire by the end of 2018, per its contract with the state, or even the 260 needed in order to receive subsidies, an audit found the company had managed to hire only 113. At the Mount Pleasant campus, it had erected a single structure, a 120,000-square-foot space that sat virtually empty. Its very name, “the multi-purpose building,” seemed noncommittal. As for the promised LCD factory, the “Fab,” Foxconn boasted in a letter that a contractor had moved 4 million cubic yards of dirt. As 2018 came to an end, the company froze budgets and canceled planned career fairs. The project entered a complete stall.
The problem, employees soon learned, was that Foxconn’s planning did not extend far beyond the broad promises made to Wisconsin officials: an enormous LCD plant, some other manufacturing, lots of jobs. There appeared to be no research into the market for products Foxconn might make or the costs of producing those products in Wisconsin. Employees know this because many of them, no matter what role they had been hired for, were told to figure out what Foxconn should do in Wisconsin themselves.
“They asked me to create a business in Wisconsin, to come up with a business model, whatever I thought would make money,” said one of several engineers who was training in Taiwan when Foxconn called him home to help figure out what to do. He came to a conclusion shared by many who joined the project: “The most common misunderstanding with Foxconn is people here thought Foxconn had a strategy and a business plan when they were coming into Wisconsin. They did not. They had no plans at all.”
This state of affairs was obscured by the fact that Foxconn did seem to be doing something. It had bought buildings, moved dirt, and hired people. But this apparent progress, as well as much of the chaos that followed, can be explained by the company’s distributed power structure. While Gou has likened good leadership to dictatorship, his empire is too vast to be governed alone. The company is composed of a fluctuating number of around a dozen “business groups,” each responsible for a range of products, according to Jenny Chan, assistant professor of sociology at the Hong Kong Polytechnic University and co-author of Dying for an iPhone, which examines Foxconn working conditions following a string of employee suicides in the 2010s. These business groups have a large degree of autonomy and are responsible for their own profitability, and are themselves made up of smaller units that also operate quasi-independently, even doing business and competing with other Foxconn entities.
Much of what had looked like progress in Wisconsin had been done by a Foxconn subsidiary called Flying Eagle Wisconsin, or FEWI, which had been created to lay the groundwork for the business groups that were supposed to actually manufacture things. And when it came time for Foxconn’s other business groups to invest in Wisconsin, they balked.
The first half of 2019 was, in the words of one employee, a “shoving match” between Gou and Foxconn’s business leaders over committing to what employees say seemed to be viewed as a money-losing debacle. “I was told on my first day, the only reason my business unit was there is because Terry Guo told them to go,” said an employee. Another employee was told by their supervisor that they were only in Wisconsin because the Trump administration wanted them to be, while three others recall a meeting where Woo, Gou’s special assistant, said that if it were up to him, Foxconn would immediately leave the state, but that they had to make Gou’s commitment work.
At one point, people were stuck without desks at the Milwaukee headquarters because the Foxconn subsidiaries they worked for refused to pay rent to FEWI, according to one employee. Recruiters say the Foxconn subsidiary charged with LCD manufacturing was slow to hire and seemed uninterested in moving the project forward. “They were never really there,” said one recruiter. “Then it just kind of fizzled.”
Foxconn’s vacillations spilled into public view in January 2019, when Woo told Reuters, “In Wisconsin we’re not building a factory,” having finally discovered it was unprofitable to make LCDs in the US. The comment caused an uproar. State Republicans swiftly blamed Evers for driving Foxconn out; the administration expressed surprise at the change; Trump spoke with Gou, and Foxconn immediately announced that LCD production was back on. “Great news on Foxconn in Wisconsin after my conversation with Terry Gou!” Trump tweeted, claiming credit for bringing Foxconn to Wisconsin a second time.
If the factory was meant to earn Trump’s goodwill, the January incident showed that the company couldn’t simply vanish as it had elsewhere. Foxconn was stuck in Wisconsin, and it needed to find a way to cut its losses. Employees at every level of the project were enlisted in a search for something — anything — Foxconn could do to generate revenue.
Alan Yeung, Foxconn’s director of US strategic initiatives, had a head start. As chief of FEWI, he had relatively free rein before budgets were frozen and Woo took on a greater role. But Yeung, employees say, had always been more interested in various eccentric side projects than manufacturing. Chief among them were the “innovation centers.” An innovation center, Yeung explained, would “help inspire innovative ideas,” “catalyze cutting-edge solutions,” and “play a key role in building a vibrant AI 8K+5G ecosystem.”
In reality, the “innovation centers” were meant to be co-working spaces. “Alan constantly referenced WeWork and how we should be following that model, thinking it was brilliant,” said an employee. It was branded “Blaze.”
But WeWork, as WeWork itself would soon prove, is a difficult business model to make profitable, particularly with buildings that required major renovations and were located in far-flung corners of Wisconsin. Foxconn did not approve funds to remodel the buildings. Still, Yeung persisted in trying to make Blaze happen. He solicited designs and awarded bids to contractors, who then sat for months waiting for instructions. (One is still waiting to be told what to do with an HVAC unit he bought more than a year ago.) He hired a Blaze “concierge,” who, because Blaze did not exist, simply sat in the Milwaukee headquarters with nothing to do. The buildings remained empty, save for their original tenants — banks, an architecture firm, other small businesses — which made them, ironically, one of the only profitable ventures Foxconn had in the state.
With Blaze stalled, employees began convening to discuss literally any other idea to make money. They searched for things in Wisconsin they could export to China: cosmetics, designer handbags, ice cream, carp. Yeung asked them to draw up a plan for building an aquaponic fish farm in Mount Pleasant, having been inspired by a company in northern Wisconsin and reasoning that Foxconn had access to cheap water the state provided for LCD manufacturing. They briefly explored doing something with esports, maybe sponsoring a gaming team that could use the empty innovation centers, according to one source. A plan to export dairy to China got as far as a meeting with the Wisconsin Department of Agriculture before collapsing. The department confirmed it met with Foxconn representatives about the plan in the spring of 2019 but said Foxconn paused “to conduct more research” and no additional discussions have taken place.
Yeung referred a request for comment to a Foxconn media representative, who did not respond.
Yeung’s other major FEWI initiative was something called “smart cities,” and for a time, it seemed to be getting traction. Yeung had found an eager audience for his smart city pitch in Cory Mason, the Democratic mayor of Racine, a deindustrialized city with high unemployment near Foxconn’s Mount Pleasant campus. At an October 2018 announcement of a smart city-focused “innovation center” in Racine, Mason had invoked the city’s history as a manufacturing powerhouse and predicted a return of past glory. “I think people are going to be very surprised and impressed with what we have ahead of us,” he said.
In meetings at Racine’s City Hall, Foxconn representatives and city officials started developing a plan, elements of which Racine submitted to a competition called the Smart Cities Readiness Challenge in 2019: camera-festooned autonomous vehicles would patrol high-crime areas, the city said in its proposal, guided by 5G cells mounted on lamp posts. Self-driving vehicles — retrofitted golf carts at first, then shuttles as soon as 2020 — would ferry Racine’s workers to Foxconn’s campus. Foxconn, the city noted in the submission, was a “particularly important stakeholder” and would help provide financing and technology.
In March 2019, Mason and Foxconn’s Woo released a memorandum of understanding to make the vision a reality, and the next month, Racine won the competition, the smallest city ever to do so.
But when city officials started asking basic questions about the sort of infrastructure they needed to build in order to accommodate Foxconn’s technology, Foxconn employees found they were unable to get clear answers from the company. “They were losing confidence, and then we parade in more new shiny ponies, and more people who couldn’t answer what should have been easy questions,” an employee said.
“The whole thing was a facade,” said another employee. “It was an idea that Alan baked up and tried to commit the business to. The business eventually said no, it doesn’t make sense, it costs too much money. Poor Racine was left scratching their head and saying, ‘What just happened?’”
Shannon Powell, Mason’s communications director, said Woo later told officials that Foxconn had to focus on its manufacturing project in Mount Pleasant and that the Smart City project was being put on indefinite hold. The city appears to have moved on: last September, Racine announced a new Smart City initiative — with US Cellular.
Foxconn only ever got as far as buying the golf carts. They arrived from China disassembled, in orange, pink, and other festive colors. One employee described them as “the biggest pieces of shit,” like something “bought off Wish.com.” Unable to make them autonomous, Foxconn put them in storage in the multipurpose building. At one point, the company discussed outfitting them with lights and turning them into security vehicles, but the subsidiary in charge of security refused to pay FEWI for the carts, according to one employee. As the divisions bickered, bored employees would come down from the Milwaukee headquarters to race the carts around the empty building, until the batteries finally died.
As Foxconn was secretly careening from idea to idea, Wisconsin Republicans ferociously insisted that the project was on track and did their best to derail any attempted oversight from the Evers administration. In April 2019, Vos swiftly attacked Evers after the governor mentioned that the contract with Foxconn would have to be revised, given that the company admitted it had changed plans from a Gen 10.5 factory to a Gen 6, with a likely reduced number of employees. “It is beyond my level of understanding to think that a governor of Wisconsin is basically rooting for the failure of the largest economic development in our state’s history,” Vos said, adding that he had met with Foxconn more times than he could count, and that “their goal has never wavered from creating 13,000 jobs.” Foxconn’s Yeung tweeted that Evers’ statement about the contract revision was “probably fake news,” adding a laugh-crying emoji.
It turned out that revising the contract had been Foxconn’s idea. Woo had proposed it to Evers the month before. Not that it mattered, because actually revising the contract would have required Foxconn to disclose its revised plans, and possibly because it had none, it refused to do so. Evers, whose control of WEDC had been limited by the Republican assembly during Walker’s lame duck period, appeared unable to penetrate the fog. Documents show an Evers administration official emailing Foxconn representatives repeatedly in June and July 2019, saying that employees in Wisconsin don’t seem to know when the factory will be finished, what it will make, or how many people will work there.
But a growing number of people who had tried to work with the company were starting to notice something wasn’t right.
By spring 2019, business owners who had tried to become Foxconn suppliers realized they had been ghosted. The year before, WEDC had held a series of events to connect Foxconn with local suppliers. One attendee The Verge spoke with entered his business into the supply chain system and never heard back; another had the company go dark after several meetings; a third says he filled out an extensive survey only for a consultant to request approximately $10,000 to have his business vetted for Foxconn’s mysterious project. (He declined.) One explanation for this is simple: Foxconn, not knowing what it was going to make, hadn’t known what suppliers it needed. “We’d take their name, and we’d never call them again,” acknowledged a Foxconn employee.
Wisconsinites who attended the company’s enormous hiring events fared little better. Especially during the late 2018 hiring spree, these events could have what employees describe as a “speed-dating”-like atmosphere, with applicants offered jobs on the spot and given 24 hours to accept. The problem was that the Wisconsin project never had a conventional budget: everything from printing business cards to hiring people had to be sent back to Taiwan for approval, a process that could last months and which often ended in a denial. (Precisely who had power to approve budgets was a mysterious and always shifting matter for the Wisconsin employees; for a time, they say expenses required the approval of a figure referred to exclusively as “Money Mama.”) This put recruiters in the position of hiring for jobs they thought were real, only to delay start dates and sometimes rescind offers entirely. “They destroyed a lot of lives with false hope and promise for something that never happened,” said an employee.
“It was such a charade, looking back on the whole thing,” said Chris, who attended a career fair in Mount Pleasant in late 2018 and asked that only his first name be used. Foxconn had heavily advertised its recruiting of military veterans like himself, and he was excited “as a lifelong cheesehead” about the prospect of Wisconsin becoming a tech hub. So he was thrilled when, after an interview with Yeung, he was handed a letter, viewed by The Verge, declaring the company’s intent to hire him in two weeks. He left his current job — he had been planning to anyway — and waited. Two weeks came and went. He followed up and was told Foxconn was just trying to find the perfect role for him. Then he was told he would be placed on a different team, but the team leaders were in China and would be back in two weeks. Then he was told to come to Milwaukee to meet them. Then Foxconn canceled. Eventually, he found another job.
“I just realized there was no there there,” Chris said. “Now whenever I see them in the news, I laugh and flip the channel, because I just don’t believe any of it. I just wonder, at the end of the day, what was their endgame? What did Foxconn get out of it? Why the elaborate statewide charade?”
Many Wisconsin institutions had reason to ask that question. Earlier partnerships announced with local companies like Rockwell Automation had been followed by total silence. (Employees say they quickly fizzled; Rockwell did not return a request for comment.) Of the $100 million gift Gou promised the University of Wisconsin-Madison, the school confirmed that only $700,000 ever arrived.
Meanwhile, employees dispatched in search of business ideas had worked out deals and partnerships with local companies only to have Foxconn leadership not move forward. Eventually, they started running into bridges that prior Foxconn delegations had already burned. “Every place we tried to go, somebody from Foxconn had already been there and they’d already pissed them off,” said an employee.
Foxconn, according to two employees, turned to one of the few remaining friendly prospects: the Trump administration. Woo flew to DC to meet with Trump trade adviser Peter Navarro, who suggested Foxconn become a vendor to the US General Services Administration (GSA), the agency that procures products for the federal government, according to an employee with knowledge of the visit. Foxconn named the initiative Project Red, White, and Blue.
Navarro did not respond to a request for comment, but on July 12th, 2019, he told CNBC that Foxconn representatives had visited his office and that “the White House is actively engaged” with the company and helping it overcome any obstacles it might face. (He also disputed the suggestion that the project had been scaled down and said Evers was “not a friend of ours, and he’s not a friend of the project.”)
The solution had a neat circularity to it — a government-subsidized project meant to earn government goodwill finding salvation in selling back to the government — but like almost everything Foxconn tried, it ended in failure. Employees searched for products Foxconn already made that were eligible for the GSA, but everything was either unprofitable or already being sold by major Foxconn customers like Dell and Cisco, according to a source with knowledge of the project, which would mean Foxconn would risk undermining actually profitable business relationships to salvage the Wisconsin project.
Increasingly desperate, Foxconn cast about for companies willing to move into the industrial park. Employees looked for local manufacturers that might rent space in the “Fab,” restaurants and cafes that might want to use its real estate, even companies that wanted to use its abundant land to store construction equipment, snowmobiles, or boats.
Every plan seemed to stall or get vetoed by Foxconn leadership. “All we did was create slide decks to get beat up by executives,” said an employee. “It was extremely frustrating,” said another. “I left a very good job to come there.”
In May, Gou returned to Wisconsin. Meeting with Evers on his way back from another visit to the White House, the founder, now running for president of Taiwan, said Foxconn’s plans had “never changed” and that LCD production would begin in May 2020. Evers appeared to backtrack on his previous criticism, saying he didn’t doubt Foxconn’s word.
But employees saw that yet another pivot was afoot. They were told to gather in Mount Pleasant, and Gou introduced a new face: Jay Lee, a University of Cincinnati professor and vice chairman of Foxconn Industrial Internet (Fii), a Foxconn company that makes networking equipment. The third leader in less than a year had been dispatched to Wisconsin.
Employees greeted Fii’s arrival with cautious optimism. Fii was ascendant in the Foxconn empire, having spun off from the main company and gone public in China in 2018, and its focus on automation made it less subject to the higher labor cost of operating in the US. Most importantly, Fii had money: it finally authorized funds to repair the Milwaukee office’s temperamental elevators. In September, Fii announced the first manufacturing deal of the Foxconn project, a temporary contract to make coffee vending machines for a company called Briggo.
But Fii also brought a cultural shift to Foxconn in Wisconsin. Employees say that while Fii’s Jay Lee was preoccupied with starting a vague “AI Institute” and giving talks on automation, the real power lay with Fii CEO Brand Cheng. In interviews, employees referred to Cheng as “the tyrant,” “the bully,” and “mini-Terry,” for a managerial style he took from Gou, who himself cites Genghis Khan as a personal hero and whose book of quotations, which workers at Foxconn’s Chinese factories sometimes have to transcribe as punishment, includes aphorisms like “growth, thy name is suffering.”
“My god, the incessant screaming, it was like a Saturday Night Live skit,” recalled one employee. Two employees witnessed and others heard about an incident in which Cheng, leaving a meeting, saw the elevator open and began berating an employee for having failed to replace its carpeting, which Cheng found ugly. The employee, an early hire and a veteran, protested that he had only been using the byzantine expense approval process Cheng himself had demanded they follow; this process, employees say, seemed designed to not pay bills, and meant the nominal purchase of new carpet squares had to be sent overseas for approval.
“Brand just blew up, and said, ‘This is ridiculous, I could do it myself in a couple of hours, you don’t know what you’re doing, why are you here?’” one employee recalled. “Here’s a Navy officer being publicly embarrassed like that, and everybody felt horrible for him. And the next thing is ‘oh, shit, when’s it going to happen to me?”
The employee quit shortly afterward, they said. (He did not respond to inquiries.) Afterward, others went out and bought the carpet themselves. “It was done just so Brand would stop screaming,” said an employee. “People were using a lot of their own money just to avoid bad situations.”
“I thought that was mild,” another employee said of the elevator incident. What stood out to him and another employee was a meeting in which Cheng, yelling at a co-worker, made his hand into the shape of a gun. “He said, ‘If you don’t get your job done, I’ll get somebody who can, it’ll be just like this,” the employee recalled. “He put it right to the guy’s temple and pulled the trigger.”
But everyone knew Cheng was hesitant to actually fire anyone because Foxconn still had to hit the subsidy threshold, so all there was to do was scream. Even Yeung came in for it, employees said, with Cheng making belittling comments about his bad business sense. The Wisconsin employees, accustomed to the more polite environments of Midwestern offices, were appalled by these outbursts. They learned to sit along the walls in meetings, far from Cheng; when Cheng was overseas, they jockeyed for seats out of the webcam’s frame. Someone from HR tailed Cheng around the office in an apparent attempt to intervene before disaster, one employee said. The new management seemed increasingly open in its disdain for the local workforce. “Why are Americans paid so much and do so little? I can’t tell you how many times we heard that,” said one employee. “It was certainly a toxic work environment,” said another. “There’s no reason you can’t be nice and respectful.”
Neither Cheng nor Lee returned requests for comment.
Perhaps sensing all was not well, Lee ordered everyone to watch the documentary American Factory, about a Chinese automotive glass company that buys a shuttered GM plant in Dayton, Ohio. But employees watched the film — in which hopes for a manufacturing resurgence in the Midwest founder amid culture clashes and differing labor standards, Chinese supervisors explain that Americans are lazy and need to be flattered (“Donkeys like being touched in the direction their hair grows”), and in which American management is eventually fired and replaced with Chinese leadership — with the dawning horror of recognition.
“We were like, ‘Oh my god.’ That’s exactly what’s occurring at Foxconn,” said an employee. “This is not a good thing. This is bad! But he couldn’t tell.”
It wasn’t long before Fii encountered the same lack of direction and reluctance to invest that had beset the earlier iterations of the Foxconn project. In September, Fii submitted plans for two buildings: a 261,000-square-foot manufacturing center and a nine-story glass sphere the company said would be a network operations center for an adjacent server farm, only to withdraw the sphere plans later the same day.
The original plan had been grandiose: the sphere was to be the dot in the “i” of a complex of data centers spelling out “Fii” when viewed from the air. But according to three employees, Foxconn balked at the cost. An employee with knowledge of the project said that Foxconn finally moved forward with the sphere — and only the sphere — when the architect told the company it had to put a deposit down for the steel if construction was going to finish in time for a long-promised visit from Trump.
But the building without the data centers was just a glass orb in a field — at best, “really, really, really expensive office space,” in the words of one employee. Adding to Fii’s troubles, FEWI, also trying to cut its losses, had “tricked” Fii into buying more land than was needed for the sphere, according to a second employee. A Foxconn executive briefly entertained an elegant solution, according to two employees: starting a Foxconn tree farm, so the company could get free trees for the terrarium-like interior of the sphere that Gou wanted, and sell the excess trees for profit.
“It’s endless,” said an employee, noting with frantic exasperation that the sorts of tropical trees Foxconn wanted can’t even grow in Wisconsin’s climate. “When you’re desperate and you have no product to sell and the only asset you have is land, what can you do? You build on it or you grow crops on it.”
Hopes that Fii would right the ship were fading. There were the same shifting plans, the same reluctance to spend money. Fii itself was partly running on IT equipment an employee brought in from home, according to an employee, and the Mount Pleasant facilities had its internet temporarily shut off over unpaid bills, according to two others.
It was the same as it always was, but now with more yelling.
There were always two hard deadlines under Fii: the Trump visit and the cutoff for hiring enough people to get subsidies. The Trump visit was codenamed the “505 project” because it was originally scheduled for May 5th, 2020, before Foxconn’s delays and then COVID-19 turned it into the “909 project,” then postponed it indefinitely.
“There was supposed to be a big announcement and a ribbon cutting that would coincide with Trump coming to town, which would also coincide with his campaign trail,” said an employee. “The joke there was, what’s he going to be cutting the ribbon to?” There was talk of reviving the still-not-self-driving Smart City golf carts. Employees joked about having someone stationed behind a curtain, steering the president around via remote control.
Some employees hoped that a visit from the president would give Foxconn the impetus it needed to finally figure out a business and begin manufacturing. But as the date approached, the event became more of a trade show. With the “Fab” empty, and certainly not making LCDs, the plan became to use the smaller multipurpose building for the event.
A demonstration manufacturing line making circuit boards was to be set up, along with kiosks from tech companies that, while not currently buying products from Foxconn’s Wisconsin operation, theoretically might have done so in the future. In lieu of the data center Foxconn had scrapped, a shipping container-sized modular data center was to be brought in to sit beside the hastily constructed sphere, an odd disco ball-like rendition of the sleek domes built elsewhere by Amazon and Apple. It would be even less real than the groundbreaking Trump presided over more than two years before — as one employee pointed out, at least they sold some of the TVs made on the assembly line set up for that event.
“We were going to have a fake data center there, and all these brands that we don’t have relationships with, and we were going to plead that we were doing business with them,” said another employee involved in the planning. “Oh my god, that was distasteful.”
But it was the other goal, the subsidy hiring quota, that would come to consume Fii and ultimately crush what few remaining hopes Wisconsin employees had for the project.
As Vos and other fierce defenders of the Foxconn deal often point out, the state’s contract is performance-based. The company only gets paid subsidies from the state if it employs a certain number of people each year. But this mechanism also creates a perverse incentive: because employees are only counted at the end of the year and their pay is annualized to see if it meets the minimum yearly salary, Foxconn could hire people in the final weeks of the year, get tens of millions in subsidies, and be free to lay them off once the quota filing deadline passed. (In an interview with The Verge, current Wisconsin Department of Administration Secretary Joel Brennan called this “one of the real weaknesses of the contract” and indicated it was something the state would like to revise.) Meeting the hiring quotas was Foxconn’s best chance at recouping some of the losses incurred after two years spent flailing around Wisconsin and erecting empty buildings. It had failed in 2018; it couldn’t afford to do so again.
By October 2019, the company was falling behind, and Cheng instituted a daily morning meeting. “Brand came in, and started screaming at the table and saying, your group, you’re going to hire this many, so and so is going to hire this many, even if they didn’t need them,” said an employee. “It was purely nothing more than bringing people onboard to hit a number.”
Foxconn devised two programs to meet the quota. The first, called Earn and Learn, was marketed as a way for students at nearby colleges to work at Foxconn as they finished their degrees. There was no work for them to do, but provided they badged into Foxconn buildings they would count toward the company’s employment quota. “Everyone was watching YouTube or doing homework,” said an employee.
The other initiative, called Foxconn Future Leaders, targeted recent college graduates, preferably foreign students in the US on visas, according to multiple employees. “He pointed out several times that they were cheaper than US college graduates and they’ll work harder,” an employee recalled Brand saying. There was a sense among employees familiar with the program that the need for an employer to sponsor their visa meant the recruits would be willing to relocate and work for less money. “These kids, they’re just looking to stay in the country,” said an employee.
Foxconn contracted with a recruiting service that found the students. A makeshift call center was set up, with employees hiring after cursory conversations, according to an employee who witnessed the operation.
An employee who went through the program received a message from the recruiting firm on LinkedIn. He already had a good offer in his home city on the East Coast, but in his brief phone interview, a Foxconn representative told him the company would sponsor his H1B application. He decided to take it and packed to move to Wisconsin.
When he arrived, he experienced the familiar Foxconn shock: there was no work to do. More alarming, managers said if he did not find a place in a business group in six months, he’d be let go. (The offer letter, viewed by The Verge, makes no mention of this, but other employees likened survival at Foxconn to a game of musical chairs in which they had to politick their way into whichever business group seemed to currently have money.) “This ruined my career,” said the employee, who eventually found a new job in his home city and fled.
While the American employees were ridiculed for being lazy, the Future Leaders, many of whom were from China and India, were pushed to work overtime, sometimes without pay, according to employees. Two employees recalled an incident in which a manager at the multipurpose building, in which Foxconn was setting up a small circuit board manufacturing line, used a worker’s immigration status as leverage in demanding overtime work, an event also corroborated by screenshots of employee correspondence. “He was yelling, ‘you’re an H1B, I’m an EB2, if you don’t work hard enough, if you don’t work overtime, I can get you back to Mainland China anytime,’” said an employee familiar with the incident.
An employee involved with recruiting estimates that by mid-October, 90 percent of the offers he dealt with were for visa holders who would need sponsorship. Department of Labor data shows Foxconn’s Wisconsin operation filed 82 precursor applications for H1B visas in the final months of 2019 and the first half of 2020.
By December, Foxconn had hired so many students and graduates it had to put them in the multipurpose building and a nearby rented space, where they sat at folding card tables watching cooking videos and doing schoolwork. In the Milwaukee headquarters, cubicles were torn out and replaced with long tables where people sat shoulder to shoulder. “You needed a plunger to force people in the building, it was so packed with people,” said an employee.
The program, from Foxconn’s perspective, was a success. In December, Gou returned to Wisconsin following yet another visit to the White House, where he said Trump told him to “bring more jobs to Wisconsin.” Lee claimed they had already done so, boasting that by the end of the year, Foxconn would have “well beyond” the 520 employees they need to receive subsidies.
According to the company’s subsidy application, it employed 580 people at the end of 2019. Sixty percent were hired in the final two months.
The layoffs began with the new year. Starting in January, when the promised LCD factory was originally scheduled to open, Foxconn instead began letting employees go in batches.
“It was a pump and dump,” said an employee. Many of the Wisconsin residents were laid off, as were many of the local college students. Some of Foxconn’s first hires had already left, including Tank Murdoch, who held a golden shovel alongside Trump at the groundbreaking ceremony, and Peter Buck, who often served as the public face of the project and employees say quit in frustration. (Neither returned requests for comment.) An employee involved with recruiting estimated in June that Foxconn’s ranks had fallen by approximately half from its end-of-year peak, with foreign employees making up a large share of those who remained. “Mandarin is spoken more than English on my floor now,” the employee said.
Gou returned in February, and there were rumors a White House representative would attend. The ranks of local workers from Wisconsin had diminished to the point where, according to two employees, Asian employees were told to stay out of an employee photo with Gou “to show diversity in the company.”
In many ways, the Foxconn debacle in Wisconsin is the physical manifestation of the alternate reality that has defined the Trump administration. Trump promised to bring back manufacturing, found a billionaire eager to play along, and now for three years the people of Wisconsin have been told to expect an LCD factory that plainly is not there. Into the gap between appearance and reality fell people’s jobs, homes, and livelihoods.
The buildings Foxconn has erected are largely empty. The sphere has no clear purpose. The innovation centers are still vacant. The heart of the project, the million-square-foot “Fab,” is just a shell. In what an employee says was a final cost-cutting measure, only the portion that was to host the Trump visit was ever finished. Recent documents show the “Fab,” once intended for use as manufacturing, has been reclassified as a massive storage facility.
WEDC, as part of its audit of the company’s 2019 subsidy application, had Foxconn survey its employees about what they were working on. Not a single respondent mentioned LCDs because no one is working on LCDs, and they never were.
The project has fallen orders of magnitude short of its hiring and investment targets. WEDC found Foxconn had only 281 eligible employees at the end of 2019, 13 percent of what it had originally aimed for. (Many of the employees Foxconn tried to claim were paid too little or hired too late in the year to get a paycheck in 2019.) After this year’s layoffs, it is nowhere near meeting its 2020 target of 5,200 employees. Foxconn itself acknowledged, in its subsidy submission, that it has so far invested 2.8 percent of the $10 billion it promised. It has built less than 2 percent of the 20 million square feet of manufacturing space it originally planned.
The state is at an impasse. It can deny further subsidies, as it did last week, but taxpayer money has already been spent. The Village of Mount Pleasant took on hundreds of millions in debt that was supposed to be repaid by property taxes on Foxconn’s fantastic campus. It will now have to hope Foxconn honors its contractual obligation to make up the shortfall. (In a joint statement following the state’s subsidy rejection, Mount Pleasant and Racine officials said they were “disappointed in the State’s decision to not support the creation of the Gen 6 facility,” that they observe “significant construction activity” in the Village, and that Foxconn “continues to fulfill its financial obligations under the local contract.”)
“The state has upheld our part of the contract,” said DOA’s Joel Brennan. “Unfortunately, Foxconn hasn’t been able to do that on their side.” Brennan and WEDC officials continue to say they’re open to revising the contract to reflect whatever it is that Foxconn is now doing. Employees say that the company is now making servers, and Foxconn has announced it is making ventilators, though no employees or state officials could say whether any have been produced. Yet, Foxconn has so far refused to amend the contract. The company continues to insist, against all evidence, that an LCD factory is on its way.
This steadfast rejection of reality is where the Foxconn debacle stands apart from other development projects that fall short of their hype. The company’s desperate quest to maintain appearances caused it to fail repeatedly and in ways more destructive than mere ordinary failure would have been: local businesses were strung along, civil servants spent years figuring out what the company is doing, residents were removed from land the company didn’t need, and again and again recruits were lured in by the vision of a grand manufacturing renaissance in Wisconsin.
That vision got Gou regular access to the White House during a trade war and gave Trump a groundbreaking and almost a ribbon-cutting, too. But maintaining the mirage required a culture of secrecy. Employees were warned not to talk to the press (including, specifically, me). Many were afraid to speak — afraid of getting fired, or of retribution even after they’d left. Publicly, the company issued announcement after announcement — innovation centers, career fairs, smart cities, AI 8K+5G, the AI Institute — each one erasing the memory of the last missed deadline. (One employee quipped that one of the few things Foxconn succeeded in making in Wisconsin was press releases.) The illusion was defended by GOP officials at all levels of government, from Mount Pleasant to the State Assembly to the White House, who accused anyone pointing out that the project was off track of trying to scuttle it for partisan ends, as if the existence of the factory were open to debate and positive thinking might make it real.
But in actual reality, the project has succeeded in manufacturing mostly this: an endless supply of wonderful things for the President to promise his supporters. This past weekend, in an interview with a local Wisconsin TV station, Trump insisted Foxconn had built “one of the most incredible plants I’ve ever seen” in Mount Pleasant and would keep its promises and more if he was reelected. “They will do what I tell them to do,” he said. “If we win the election, Foxconn is going to come into our country with money like no other company has come into our country.”
No one wanted to believe promises like this more than the people who went to work for Foxconn in Wisconsin. They each came to regret different things: the wasted time, the jobs they’d left, the integrity lost making deals and offering jobs only to have the company change course. But one common theme was frustration that it hadn’t turned out to be real and that long after they’d learned the truth, they saw the facade still standing. “There are a lot of good people who fell for this,” said one employee, shortly before departing a job at Foxconn. “Who wanted to see it succeed.”
“We got screwed,” another former employee said. “The state got screwed.”
Design by William Joel / The Verge
Lead image photography by Scott Olson, Andy Manis, Brendan Smialowski / AFP, Billy H.C. Kwok / Bloomberg via Getty Images
Dodgers favored over Rays to win World Series
Cam after loss: ‘No need to press panic button’
Support Indie Bookstores With Your Holiday Shopping
Check out some wonderful Playdate game demos, including a low-fi Doom
Puerto Rican Piñon
Still no first stimulus check? How to track it and report your absent payment to the IRS – CNET
Tech2 months ago
Check out some wonderful Playdate game demos, including a low-fi Doom
Food2 weeks ago
Puerto Rican Piñon
Tech2 months ago
Still no first stimulus check? How to track it and report your absent payment to the IRS – CNET
Science2 months ago
Elon Musk promises demo of a working Neuralink device on Friday
Tech4 weeks ago
Fitbit Sense review: enough bugs to raise your heart rate
Tech2 months ago
Spotify Duo vs. Family vs. Individual: Which Premium Spotify plan is best? – CNET
Entertainment1 month ago
Kim Kardashian Appears To Give Birth To Kylie Jenner In Leaked Kanye West & Tyga Video — Watch
Entertainment2 months ago
Jake Paul’s LA Mansion Raided By FBI During Mysterious Early Morning Search – Watch