Category: UNCATEGORIZED

President Donald Trump and the Office of the U.S. Trade Representative have issued technology companies some temporary tariff relief.

Citing an unwillingness to hit consumers with higher prices on things like computers, mobile phones, laptops, video game consoles, computer monitors, clothes and shoes before the holidays, the President and his trade reps are holding off on slapping additional tariffs on those products coming from China.

The President could also have been motivated by growing concerns that the ongoing trade war could trigger a global recession and hurt his chances for re-election in 2020.

Whatever the reason, the news sparked a stock market rally on Tuesday with investors ignoring the rising prices that 10% tariffs on imports that don’t include consumer goods would cause.

The Dow Jones Industrial Average and S&P 500 indices were both up 1.4% on the day, while the Nasdaq rose 1.9% — thanks in large part to a surge of Apple stock. The company’s stock rose $8.49 or over 4.2% to close at $208.97.

At the beginning of the month, President Trump said he would slap a 10% tariff on $300 billion worth of Chinese goods, which sent markets tumbling. An ensuing slight devaluation of the Chinese currency further pushed markets into a tailspin before they began to recover.

The news on Tuesday all but erased those earlier losses.

At a press event in San Francisco, Twitter Product Lead Kayvon Beykpour talked about a number of product changes coming to the company’s service, he also addressed the oft-memed user request for an edit button. Long story, short, you shouldn’t expect to see the button anytime soon.

“Honestly, it’s a feature that I think we should build at some point, but it’s not anywhere near the top of our priorities,” Beykpour said. “That’s the honest answer.”

The executive said that there were some obvious risk factors but that he felt the company would eventually be able to build a feature to address user needs like correcting a typo or clarifying what they meant to say.

Twitter announced earlier in the event that the company is testing the ability to let users follow topics the same way they would ordinarily follow accounts.

As part of new efforts towards accessibility, Google announced Project Euphonia at I/O in May: An attempt to make speech recognition capable of understanding people with non-standard speaking voices or impediments. The company has just published a post and its paper explaining some of the AI work enabling the new capability.

The problem is simple to observe: The speaking voices of those with motor impairments, such as those produced by degenerative diseases like amyotrophic lateral sclerosis (ALS), simply are not understood by existing natural language processing systems.

You can see it in action in the following video of Google research scientist Dimitri Kanevsky, who himself has impaired speech, attempting to interact with one of the company’s own products (and eventually doing so with the help of related work Parrotron):

The research team describes it as following:

ASR [automatic speech recognition] systems are most often trained from ‘typical’ speech, which means that underrepresented groups, such as those with speech impairments or heavy accents, don’t experience the same degree of utility.

…Current state-of-the-art ASR models can yield high word error rates (WER) for speakers with only a moderate speech impairment from ALS, effectively barring access to ASR reliant technologies.

It’s notable that they at least partly blame the training set. That’s one of those implicit biases we find in AI models that can lead to high error rates in other places, like facial recognition or even noticing that a person is present. While failing to include major groups like people with dark skin isn’t a mistake comparable in scale to building a system not inclusive of those with impacted speech, they can both be addressed by more inclusive source data.

For Google’s researchers, that meant collecting dozens of hours of spoken audio from people with ALS. As you might expect, each person is affected differently by their condition, so accommodating the effects of the disease is not the same process as accommodating, say, a merely uncommon accent.

A standard voice-recognition model was used as a baseline, then tweaked in a few experimental ways, training it on the new audio. This alone reduced word error rates drastically, and did so with relatively little change to the original model, meaning there’s less need for heavy computation when adjusting to a new voice.

The researchers found that the model, when it is still confused by a given phoneme (that’s an individual speech sound like an e or f), has two kinds of errors. First, there’s the fact that it doesn’t recognize the phoneme for what was intended, and thus not recognizing the word. And second, the model has to guess at what phoneme the speaker did intend, and might choose the wrong one in cases where two or more words sound roughly similar.

The second error in particular is one that can be handled intelligently. Perhaps you say “I’m going back inside the house,” and the system fails to recognize the “b” in back and the “h” in house; it’s not equally likely that you intended to say “I’m going tack inside the mouse.” The AI system may be able to use what it knows of human language — and of your own voice or the contest in which you’re speaking — to fill in the gaps intelligently.

But that’s left to future research. For now you can read the team’s work so far in the paper “Personalizing ASR for Dysarthric and Accented Speech with Limited Data,” due to be presented at the Interspeech conference in Austria next month.

You may soon be able to organize Twitter’s web of hashtags and handles in a smarter way, that is if the company can pull off its ambitious new rethinking of the app’s timelines.

The company isn’t getting rid of the process of a following users but at a press event in SF, company execs announced that they are planning to push users to start following “topics” that bring in well-engaged tweets from a variety of accounts that the user might not necessarily follow. Twitter is currently testing the feature on Android with topics focused around sports “from MMA to Formula 1” to specific professional franchises.

The company plans to greatly expand the scope of these topics so that fans will be able to have timelines devoted to BTS and skincare routines.

The company is curating the overall topics manually, but Twitter will be relying on machine learning to intelligently populate the topics themselves so that the tweets can stay up to date. The company is also testing the ability to not only follow topics in your central timeline, but create your own secondary timelines that you can bring multiple topics, accounts and hashtags into.

A feature that Twitter says it is also starting to experiment with is the ability to temporarily unfollow a topic so you can keep certain tweets out of your timeline like tweets chronicling an ongoing finale of a TV show or a football game.

Hello and welcome back to Equity, TechCrunch’s venture capital-focused podcast, where we unpack the numbers behind the headlines.

This week was a bit special. Instead of meeting up at the TechCrunch HQ to record the episode, Kate and Alex met up in muggy Boston at Drift’s office, where we linked up with Axios’s Dan Primack. And since we were feeling chatty, we went a bit long.

After checking in with Primack (he has a newsletter and a podcast), we first dealt with the latest from Tumblr. In short, Verizon Media is selling Tumblr to Automattic for a few dollars. How did Verizon wind up owning Tumblr? Ah. Well, Yahoo bought it. Later, after Verizon bought AOL, it bought Yahoo. Then it smushed them together and called it Oath. Then Verizon decided that it didn’t like that much and renamed the group Verizon Media. But Verizon doesn’t want to own media (besides TechCrunch, of course), so it sold Tumblr to Automattic, a venture-backed company best known for operating WordPress.

That’s a lot, I know. What matters is that Yahoo bought Tumblr for more than $1 billion. Verizon sold it for around $3 million. Now, Automattic now has a few hundred new employees and a shot at juicing its userbase before it goes public.

After that, we lamented that the WeWork S-1 had yet to appear. This was a tragedy, frankly. We had expected to spend half the show riffing on WeWork’s financials, alas…

So we turned to some normal material, like Ramp’s recent $7 million raise to take on Brex, and, SmartNews’s recent round, which gave it an eye-popping $1.1 billion valuation.

We ran a bit long because we were having fun, fitting in some conversation surrounding the notes from the SEC regarding the now-dead and then-fraudulent Rothenberg Ventures. More on that here if you want to get angry.

And finally, Vision Fund 2. It’s been a big source of interest for everyone on the show, and we expect whatever the second-act Vision Fund winds up becoming to be a big damn deal. The fund will invest in more than just consumer marketplaces, in fact, it’s eyeing more AI businesses and even biotech. That should be interesting.

All that and we have a lot more good stuff coming. Thanks for listening to the show, and we’ll be right back.

Recharge, a startup that tried convincing hotels to let its customers rent their rooms by the hour and even minute, has revamped and rebranded. Now Globe, the company is hoping to convince guests to sign up for short stays instead in people’s homes so that they can kick back between other commitments, and, if the host allows it, to shower and nap.

It’s at once crazy sounding and intriguing, which is perhaps why the popular accelerator program Y Combinator accepted the company into its most recent class of companies. (It shows off its newest batch of startups next week.) YC was famously early to spy the opportunity that Airbnb could chase, after all. The question is whether Globe, which likens itself to an Airbnb for day breaks, will have anywhere near the same appeal.

Its proposition is certainly similar. Home owner or renter wrings out some extra income by renting out all or part of their home, except that unlike with Airbnb, where the minimum stay is at least one night, with Globe, a host rents out his or her space for smaller increments of time.

In a world where the economic divide continues to grow between the haves and have-nots, it’s easy to see the logic in maximizing an underutilized asset — even one’s living room — in order to live more comfortably. It’s especially easy to see the logic in prohibitively expensive cities like San Francisco and New York.

At the same time, letting in a stranger — even a “businessperson” — for a shorter period of time is not going to be a no brainer for many people who might otherwise rent their home while away for a weekend. And on the other side of the marketplace, getting enough hosts with nice enough places to become hosts is a high hurdle for Globe to surmount. After all, if someone is looking for alternative to Starbucks for a few hours, and that individual has to take some form of transportation to get to a host’s couch that may or may not be as nice as pictured, that individual may well go the coffee shop route instead. (The company is also up against startups like Breather that offer hourly or daily “space as a service.”)

Founder Manny Bamfo appreciates the challenge he says. In fact, after running Recharge for a couple of years, he’s gotten well-acquainted with adversity.

Though he says that Recharge wound up seeing $4 million in revenue from its hotel partners, renting rooms to Recharge customers “wasn’t their number one priority, and that made it hard to provide a consistent experience for our customers,” he says. It was “particularly difficult for [hotels] to get their unionized cleaning labor to galvanize and get behind [the concept of cleaning rooms more frequently],” which is why the company decided to relaunch as a home-sharing service instead.

It’s not just a branding exercise. Along with the new name, Globe is starting from scratch with a new cap table, though Bamfo says Globe opened up a small round for previous investors that was “oversubscribed instantly.” Recharge had raised $10 million from investors. One of these backers was Binary Capital, which has since evolved into little more than a tangle of lawsuits. Another backer was the real-estate focused firm Fifth Wall Ventures, which maintains a small stake in the new company, says Bamfo.

In the meantime, Globe is looking to “do a proper seed round at [YC’s] Demo Day.” It’s also busy spreading the word in an effort to build up its burgeoning new marketplace of homes and apartments for rent, and advertising a rate of $50 per hour to people who host their entire home by the hour and $25 per hour to those who share less room. (Globe keeps 20 percent of the fee.)

Last but not least, Globe is also promising $1 million in general liability insurance and, for now, guests who have been verified and vetted by Bamfo himself.

It’s not a scalable solution, he acknowledges, but at the moment, he says, it’s all about building the right community and he sounds optimistic — of course — about its odds.

“People view it like selling a lamp on Craigslist. ‘If it’s not much work, and it’s another form of income, I’ll do it.'” There are a “lot of people with great jobs living in cities that are very expensive — people who are cops, who are teachers, who aren’t quite making six figures, and any extra income is a godsend.”

Asked then why Airbnb isn’t already chasing the same opportunity, Bamfo says it’s basic time management, and also a different market opportunity. “For any company to do this well, it has to be their number one priority.” Besides, he adds, Airbnb is “a travel company. we’re localized, with the ability to charge on a minute-by-minute basis. It’s a huge engineering undertaking and, for now, it’s part of our moat, too.”

A group of tech leaders has banned together to speak out against white supremacy and rampant hate speech on tech platforms. The group, Build Tech We Trust, refers to itself as a collective of tech CEOs, activists, changemakers and workers who are committed to countering hate and terrorism.

In a public letter published today, Project Include CEO Ellen Pao, Code 2040 CEO Karla Monterroso, ReadySet CEO Y-Vonne Hutchinson, Project Include Founding Member Erica Baker, Block Party CEO Tracy Chou and others make a call to hold tech platforms accountable and “build tech we trust.”

Despite platitudes by tech CEOs that their respective platforms are designed to bring the world together and foster connection, these platforms too often cause harm and “are radicalizing and fragmenting communities by providing an unprecedented ability to coordinate attacks and amplify hate,” the letter states.

That’s not to say that tech companies have done nothing to try to combat hate speech and white supremacy, but what they’ve done just hasn’t been enough. In June, former ACLU Washington Director Laura Murphy said Facebook’s white supremacy policy, despite some changes, was still too narrow. Meanwhile, stories have recently emerged regarding how people become radicalized on YouTube.

The letter comes shortly after the mass shootings in El Paso, Texas and Dayton, Ohio where many of the victims were either Latinx or black. Tech leaders in the letter also note other shootings where people were targeted because of their race, sexuality and/or religion, like Pulse Nightclub shooting and Charleston church massacre.

“White supremacist terrorism and violence, fueled by racism and misogyny, and empowered by technology, is on the rise,” they write. “They’ve moved beyond their white robes and hoods to social media and public rallies where they radicalize and fund their growing membership. Our government leaders at the highest levels encourage and spread it. Our industry leaders enable and profit from it. Four of the five worst gun massacres in modern history have taken place over the past two years. Evidence shows that many of these shooters are inspired by white supremacist ideology and targeting marginalized people.”

The aim of the letter is to serve as a call to action to encourage their fellow technologists to build ethical and responsible tech platforms.

“Whether it be a walkout, refusing to build or buy tech that accelerates hate, calling out unfair anti-abuse policies that silence marginalized voices, or continuing to demand answers from those in positions of power, the time to act is now,” the leaders write.

You can read the full letter over on Build Tech We Trust. I’ve reached out to Monterroso and will update this story when I hear back.

Democratic presidential hopeful Pete Buttigieg has unveiled his plan to address the broadband gap in this country: an $80 billion “Internet For All” initiative and set of related reforms. It echoes Senator Elizabeth Warren’s (D-MA) announcement last week, which is generally speaking a good thing.

It’s detailed in a document entitled “Investing in an American Asset: Unleashing the Potential of Rural America,” which feels like it may rub people the wrong way. It seems to imply that rural America is an “asset” to the rest of America, and that its potential has not yet been unleashed. But that’s just a tone thing.

There are a number of programs in there worth looking at if you’re interested in the economy of rural areas and how it might be spurred or revitalized (for instance paying teachers better), but the internet access portion is the most relevant for tech.

Buttigieg’s main promise is to “expand access to all currently unserved and underserved communities,” including a “public option” where private companies have failed to provide coverage.

That gets broken down into a few sub-goals. First is to revamp the way we measure and track broadband access, since the current system “is inaccurate and perpetuates inequity.” It’s important this isn’t overlooked in anyone’s plan, since this is how we officially make decisions like where to spend federal dollars on connectivity.

Like Warren, Buttigieg wants to remove the impediments to public and municipal broadband options that have been put in place over the years. This will allow “community-driven broadband networks, such as public-private partnerships, rural co-ops or municipally owned broadband networks” to move forward without legal challenges. A new Broadband Incubator Office will help roll these out, and the $80 billion will help bankroll them.

Net neutrality gets a bullet point as well — “Given the FCC’s volatility on this issue, Pete believes that legislation will ultimately be necessary,” the document reads. That’s frank, and while Warren and others have spoken out in favor of an FCC solution, it is likely that legislation will eventually come around and hopefully solve the issue once and for all.

Sen. Bernie Sanders (I-VT) was the first to make net neutrality a campaign promise, though most of the candidates have expressed support for the rule in the past.

The plan is a little less specific than Warren’s, but the truth is any plan involving this amount of money and complexity is going to necessarily be a bit vague at first. Demonstrating priorities and openness to ideas and methods is the important part, as well as throwing out a giant number like $80 billion. The specifics are unlikely to see much debate until one of these people is in the Oval Office.

“To ensure greater opportunity for all, we must make a massive investment in Internet access” summarizes the Buttigieg plan pretty well. You can read the full plan here or below.

Pete Buttigieg Rural Economy by TechCrunch on Scribd

Imagine you’re a billionaire starting a new company. You’re happy to bet your entire fortune on it. As a result, capital is no constraint. How fast should you burn money?

You probably wouldn’t use the generic startup math of dividing your available capital by 18 months and burn $55.5 million a month — though it would be fun. So if capital is no longer the currency that determines how fast you go, what should?

It’s confidence, not capital, that should be the currency of acceleration at a startup — no matter if you have a million dollars or a billion dollars to burn.

Confidence is often misunderstood by those who feign it. It is not bluster or arrogance. It’s not “trusting your gut.” Competitors raising big rounds of funding shouldn’t change your level of confidence one way or the other unless they’re doing exactly what you are. Glowing press coverage helps team morale, but it shouldn’t color your assessment of readiness to scale up.

It’s also important to note that venture capital interest is a terrible proxy for founder confidence. VCs have different structural incentives than founders; in an easy money environment, placing a big bet in a hot category, backed by a good enough team, is a job well done for a VC. Remember that they have a portfolio of companies, you’ve just got the one.

So what should drive you to scale up spend? There’s no perfect answer, but if you consistently see strong customer response to your product, marketing delivering more qualified leads for less money, sales channels becoming better instrumented and more efficient, LTV expanding with product improvements and lower churn thanks to your customer success team, you’re probably in good position to accelerate investment.

Compounding successes at all levels of the business should provide data points that give you the determination to plan out a more ambitious trajectory. The requirement for confidence shouldn’t be mistaken for conservatism. Startups need to take risks in order to thrive, but they should be calculated, not capricious. There is a limited speed any company should go based on what they’ve learned to date about their market and offering.

If you have a high degree of confidence that you can turn $1 into $2, or $10, you should invest immediately. If you don’t have that confidence, you should spend time, but limited capital, to build it. Unfortunately, too many startups feel pressure to spend money based on hope, not confidence.

Authentic growth

Startups appreciate in value through growth. This isn’t just another VC mantra: even bootstrapped startups or public companies become more valuable when they grow faster. Two $10 million companies where one is growing at 80% and the other 20% will be valued very differently. Even if the slower-growth company is generating some limited cash flow and the high-growth company is burning within reason, the high-growth company will usually be worth much more.

So given that growth drives value, why shouldn’t every startup grow as quickly as it possibly can? With capital in hand, why not spend to generate more growth and therefore more value?

Capital without confidence shouldn’t dictate a startup’s acceleration.

Shattered confidence kills startups

Companies that misuse capital as the driver of acceleration cause irreparable harm to confidence. When a company over-accelerates and misses, it takes a painful amount of time to observe the mistake, admit the mistake, correct the mistake and rebuild confidence with the team and investors that you won’t repeat the mistake. Eventually, the company must undertake the inevitable process of taking a huge step back to try to rebuild that faith. This requires going much slower than a similar company that has never faltered.

If you spend a small amount of money on a pilot and it fails, you’ve helped home in on what your product should be, and you’ve not burnt any credibility with your team or investors. Spend 10 times that amount and you’ll have no more confidence in what to do next, far less credibility and a diminished balance sheet. Worst yet, the next time you want to lean in on a major initiative, the lack of confidence of key stakeholders will likely overwhelm what may well be the right decision.

Three startup currencies: Confidence, credibility and capital

Companies create value by compounding learning and therefore compounding confidence in their future. As confidence grows, companies will earn credibility inside the management team and with investors. Once you have both, it usually gets easier and easier to find the right amount of capital needed to fuel that confidence. Confidence is the most important currency, followed closely by credibility, and only then, cash. By way of contrast, driving up revenue artificially by burning capital with low return on investment is not sustainable and does not create long-term value. This will ultimately damage confidence and credibility.

You can buy confidence with capital, but it’s rate-limited and there’s no benefit to scale

Arguably, there should be little difference between the acceleration of two competitive companies that have the same amount of confidence but radically different capitalizations. If both are early-stage startups and one company has $10 million in cash and the other has $1 billion, they should spend their money with the same principle in mind: what does it cost to build confidence that our most important experiments are working?

For a company with a million dollars, this may mean hiring a single inside sales rep to test out a direct channel based on some early successes with a specific type of customer. A company with a billion dollars will likely make the mistake to open global offices to meet international demand, without first validating that they can make that single inside sales rep successful. In both cases, the confidence of the management team and their ability to execute should be driving the decision, not the available capital.

Credibility is earned, not purchased

If you spend like you’re headed to $20 million ARR and only hit $10 million ARR, your business is in a very difficult position. Not only because you sustained large losses, but also because you’ve damaged confidence in execution — team members and investors won’t believe in the company’s ability to achieve the target the next time it wants to hit the gas pedal hard.

Conversely, If you confidently hit $10 million in sales and have sight lines to $20 million, you will not struggle to raise more money to achieve your goals. The more the management team meets its goals, the more confidence grows and the pace of acceleration can be increased. Compound confidence and acceleration is boundless.

One of the biggest mistakes of the startup community, fueled by an overcapitalized venture market and an overhyped argument about winner takes all market dynamics, is the belief that capital is a weapon that will win the startup wars.

Authentic confidence is the only real winning weapon at a startup. Capital can fuel that weapon, but when used without confidence, it usually becomes a weapon of self-destruction.

Y Combinator has kept an eye on cannabis startups over the years, but it’s their latest investment that’s sure to attract the attention of both marijuana users and law enforcement.

SannTek Labs, which is launching with new funding out of Y Combinator’s latest jumbo class of startups, is building a new kind of breathalyzer that can detect blood-alcohol levels but can also determine how much cannabis a person has smoked or otherwise consumed in the past 3-4 hours.

CEO Noah Debrincat say that he wants the startup’s SannTek 315 breath testing device to help officers make stronger “evidence-based decisions” rather than only relying on unsophisticated roadside sobriety tests or blood tests which can potentially take months to get results for and can lead to false positives by detecting cannabis use that took place several days prior to the test.

The SannTek breathalyzer detects the cannabis molecules in your breath, and gives a police officer a readout that lets them know if you have the drug in your system.

“We are specifically detecting Delta-9 THC, which is the predominant psychoactive component of cannabis,” Debrincat tells TechCrunch. “We understand how that gets into your breath. We understand what it does to you and the impairing side effects. And we know that if people are driving with high concentrations of that in their system, their psycho-motor skills are seriously decreased.”

A young startup building a device that could lead to people being arrested is obviously a pretty high-stakes situation, but Debrincat says they are confident in the tech and there are certifications that they’ll need in order to get the device into law enforcement hands. “The only way that a police officer will buy this is if NHTSA, the National Highway Transportation Safety Association, puts its stamp of approval on it,” Debrincat says, noting that SannTek was already in talks with the agency.

If it’s adopted, the startup’s device will be able to be used pre-arrest to give the officer a number indication of a driver’s impairment, or as SannTek further hones their device, the breathalyzer could be used for post-arrest evidentiary testing back at the precinct in a more controlled environment.

A Canadian startup building a device for U.S. law enforcement isn’t the most popular position given many of the conversations around systemic discriminatory practices that result in higher police presences in communities of color. But Debrincat hopes that the company’s device can be part of a positive shift due to the greater objectivity it promotes and its tighter window of detection.

“The state of affairs currently is that there’s a bunch of misdemeanor charges for small weed crimes that are happening across America and one way to address that is by federally decriminalizing the drug, sure,” Debrincat says. “But what gives police offers power now is the ability to make a call because there is no [breathalyzer] device.”

There are reasons to be concerned for law enforcement getting a tool that could be used discriminately, but Debrincat says there is also plenty of reason to be concerned for the other drivers that are on the road while cannabis users might be driving impaired. The CEO tells me that drivers are 2x as likely to get into an accident while operating vehicles under the influence of cannabis. Other studies reinforce the risks of driving while high. 

The company wants to keep the price of their device low enough that precincts across the U.S. can easily afford them, right now Debrincat says the startup is shooting for the $800-$1,000 range to stay competitive with other options out there.