Tech Brew Ride Home - Tue. 03/21 – TikTok Showdown
Episode Date: March 21, 2023Ahead of a major showdown in Congress this week, TikTok wants the US to know exactly how popular it is. Why ChatGPT was shut down for a while yesterday. Why Google is suspending PinDuoDuo from the pla...y store. The AI powered doctor’s office of the future. And is Andy Jassy the David Moyes of Amazon? If you get that reference, congratulations. Down Round Podcast (Listener Ad!) Links: TikTok says it now has 150m monthly active users in the US (MusicAlly) TikTok bans deepfakes of nonpublic figures and fake endorsements in rule refresh (The Verge) Google flags apps made by popular Chinese e-commerce giant as malware (TechCrunch) OpenAI Shut Down ChatGPT to Fix Bug Exposing User Chat Titles (Bloomberg) Amazon’s post-Bezos experiment hasn’t gone exactly as planned (CNBC) OpenAI-powered app from Microsoft will instantly transcribe patient notes during doctor visits (CNBC) Learn more about your ad choices. Visit megaphone.fm/adchoices
Transcript
Discussion (0)
On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco.
Hey, who did this to you?
What happened next turned the story into a political firestorm.
Reports have identified the victim as Bob Lee, the founder of Cash App.
From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16.
Welcome to the TechMeme right home for Tuesday, March 21st, 2023.
I'm Brian McCullough today.
Ahead of a major showdown in Congress this week, TikTok wants the U.S. to know exactly how popular it is.
Why ChatGPT was shut down for a while yesterday, why Google is suspending PIN Duo Duo from the Play Store,
the AI-powered Doctor's Office of the Future, and is Andy Jassy the David Moyes of Amazon.
If you got that reference, congratulations.
Here's what you missed today in the world of tech.
So again, it's starting to feel like there is some sort of momentum building for some sort of
TikTok showdown with the U.S. government, either an outright ban, maybe a forced divestment,
but something.
In fact, let me put this on your radar.
A literal showdown is coming before Congress.
This Thursday, TikTok CEO, Shuzi Chu, will be appearing before the House Energy and Commerce
Committee.
And, well, TikTok seems to be willing.
to come to play. Quoting musically, as part of the run-up to the hearing, TikTok has announced a new
stat for its U.S. business. It now has 150 million monthly active users there, which is up from the
100 million figure it published in August 2020. Obviously, the announcement of the new milestone is
no coincidence. It's a pointed reminder to politicians calling for a ban on TikTok of the app's
sheer scale in the U.S. Well, more accurately, it's a pointed reminder to the politicians who haven't
yet called for a ban and might still be swayable to TikTok's cause. As U.S. Commerce Secretary
Gina Riemondo told Reuters, quote, the politician in me thinks you're going to literally
lose every voter under 35 forever if you ban TikTok, end quote. Talking of young people,
the Washington Times reported that TikTok is, quote, mobilizing a cadre of influencers to descend on
Washington for some last minute lobbying tomorrow Wednesday ahead of the hearing, end quote.
Meanwhile, TikTok refreshed its community guidelines partially to address the explosion of AI content,
expanding a section to cover synthetic and manipulated media or deepfakes, but you can probably
see this as a strategic PR move as well. They don't want a flood of headlines about fake
content on TikTok at this moment. Quoting The Verge, the bulk of these moderation policies or community
guidelines in TikTok's parlance is unchanged and unsuppressed.
There's no graphic violence allowed, no hate speech, and no overtly sexual content,
with great-aided rules for the latter based on the subject's age. One newly expanded section,
though, covers, quote, synthetic and manipulated media, end quote, aka AI deepfakes,
which have become increasingly popular on the app in recent months. Previously, TikTok's
rules on deepfakes were restricted to a single line banning content that could, quote,
mislead users by distorting the truth of events or cause significant harm to the subject of the
video, end quote. Now the company says, all realistic AI generated and edited content must be,
quote, clearly disclosed as such, either in the video caption or as an overlaid sticker.
TikTok says it will not allow any synthetic media, quote, that contains the likeness of any
real private figure, end quote, or that shows a public figure endorsing a product or violating the
app's other policies, i.e. its prohibitions on hate speech. The company defines public figures
as anyone 18 years of age or older with, quote, a significant public role, such as a government
official, politician, business leader, or celebrity, end quote. A.I. generated content has increased
in popularity on TikTok thanks largely to the wider availability of AI voice cloning tools
that make it easy to mimic someone's voice. These tools have created new subgenres of content
often focused on placing public figures like President Joe Biden and former President Donald Trump
in unexpected scenarios like transposing the president's personalities into arguments about
online gaming or Dungeons and Dragons, for example.
Other use cases are more harmful.
Many AI fakes show these same figures reading transphobic or homophobic statements and have
sometimes been confused for real footage.
TikTok's prohibition on deep fake endorsements, meanwhile, seems like a response to a specific
video that used AI to fake Joe Rogan promoting a libido booster for men.
Such videos have also spread on apps like Twitter and Instagram, end quote.
Speaking of Chinese-based apps, Google has suspended PinduO Duo's app on the Play Store over what is calling security concerns and has flagged several PinduO Duo apps as malware, alerting users who had installed them, quoting TechCrunch.
In the last couple of weeks, multiple Chinese security researchers have accused Pinduo Duo, a rising e-commerce giant that boasts almost 800 million active users of making apps for Android that contain Mousycuit.
malware designed to monitor users. Ed Fernandez, a Google spokesperson said that, quote,
off-play versions of this app that have been found to contain malware have been enforced on via
Google Play Protect, referring to apps that are not on Google Play. Effectively, Google
has set Google Play Protect, its Android security mechanism, to block users from installing these
malicious apps and warn those who have them already installed, prompting them to uninstall the apps.
Fernandez added that Google has suspended PinduO Duo Duo's
official app on the Play Store, quote, for security concerns while we continue our investigation,
end quote, requesting anonymity, a security researcher alerted TechCrunch of the claims against the
apps, and said their analysis also found that the apps were exploiting several zero-day
exploits to hack users. Pinduo duo's spokesperson Kong Ho told TechCrunch in an email that, quote,
we strongly reject the speculation and accusation by some anonymous researcher and non-conclusive
response from Google, that Pinduoduo app is malicious.
There are several apps that have been suspended from Google Play at the same time, and we find it peculiar
that TechCrunch chose to single out Pind Duo Duo, end quote.
Back to AI, OpenAI was apparently forced to temporarily shut down chat GPT yesterday after a bug
let some users see the titles of other users' chat histories.
Some chat histories are still missing even after the service came back live, quoting Bloomberg.
In OpenAI spokesperson told Bloomberg that the titles were visible,
in the user history sidebar that typically appears on the left side of the chat GPT web page.
The chatbot was temporarily disabled after the company heard these reports, the spokesperson said,
the substance of the other user's conversations was not visible. A bug in an unnamed open source
software caused the problem, the spokesperson said. The company is still investigating the precise cause.
ChatGPT came back online Monday night, though users' chat histories remained missing as of late
Monday night Pacific time. History is temporarily unavailable.
read a message in the place where a user's chat GPT queries and responses typically reside.
We're working to restore this feature as soon as possible, end quote.
In the statement Monday night, the spokesperson said chat GPT is now back online, and we are working
to bring chat history back online as well, end quote.
Amazon yesterday announced plans to lay off an additional 9,000 employees, mostly in the
AWS PXT advertising and Twitch divisions after cutting more than 18,000 jobs just back in January.
but I don't want to focus on that. I want to focus instead on this analysis of what Amazon has been like
ever since Jeff Bezos left. It's from CNBC, and they make the point that either Bezos had impeccable timing, and he left his successor holding the bag,
holding a bag of maybe something less than he would want to be holding, or else Andy Jassy, at least so far, hasn't demonstrated much of the Bezos magic.
Quoting CNBC, since Jassie officially succeeded Bezos in July of 2021, Amazon has experienced.
experience its most turbulent period since the dot-com crash. Last year marked its slowest year for
revenue growth as a public company, and Jassy has been forced to guide Amazon through a series of
cost-cutting measures that nobody predicted would be necessary when business was booming through the
COVID pandemic. Amazon shares have plunged by 44% since July 5th, 2021, Jassie's first day as CEO.
And on Monday, Jassie said the company is cutting another 9,000 jobs, adding to the 18,000
layoffs that were announced in January. While the cuts represent a small percentage of Amazon's
corporate workforce, they still represent a shocking turn for a company that was in nonstop growth
phase for the better part of 25 years. Given the uncertain economy in which we reside and the
uncertainty that exists in the near future, we have chosen to be more streamlined in our costs and
headcount, Jassy wrote in an email to employees. Much of Jassy's unfortunate circumstance can be
attributed to bad timing. Historically high inflation pushed the Federal Reserve to raise rates,
crippling growth across the U.S. tech sector. But whether it's bad luck or his own missteps or some
combination of the two, Jassy is in an unenviable position as only the second CEO in Amazon's history.
For the past year, Jassy has been trimming expenses across the company. Many unproven bets like
Amazon's scout delivery robot, a virtual tours service, care telehealth program, and a video
calling device for kids were axed. He made the decision to shudder all of its four-star pop-up and
bookstores, and earlier this year, announced Amazon would close some fresh supermarkets and go cashierless
convenience marts. Dron delivery, one of Bezos's pet projects, is struggling mightily to get off the ground,
as it two faces cost cuts. The pandemic-driven e-commerce boom pushed Amazon to double its physical
footprint between 2020 and 2022. The stock soared along with headcount, but as the economy reopened and
online sales stalled, Amazon found itself saddled with more facilities than it could effectively put to use,
and eventually moved to close, cancel, or delay the opening of many new warehouses.
Earlier this month, Amazon paused the construction of the second phase of its sprawling new campus in Arlington, Virginia, dubbed HQ2.
Other construction projects in Nashville, Tennessee, and Bellevue, Washington have also been put on hold,
in part because much of Amazon's corporate workforce has been working remotely since the pandemic.
Jassy is under immense pressure to prove he can get expenses under control,
but in order to revive the enthusiasm that Bezos drove into Amazon's culture,
he's eventually got to find new engines for growth. In its fourth quarter earnings report,
Amazon barely eeked out a profit and the company issued disappointing guidance for the first quarter
with revenue growth expected to be stuck in the mid-single digits. That's not exactly what Bezos
had in mind when he told employees in early 2021 about the coming CEO transition.
Quote, Amazon couldn't be better position for the future. Bezos wrote at the time in a letter
to staffers, we are firing on all cylinders just as the world needs us to. We have things in the
pipeline that will continue to astonish, end quote. Yes, again, either this was impeccable timing or else,
maybe it's just the classic problem of filling the shoes left behind by a legend whose tenure
will probably never be equaled. The getting David Moyes to take over Manchester United after
Sir Alex Ferguson left analogy, if you know that reference. One of the biggest examples of how all this
AI stuff could go beyond parlor tricks of, you know, creating any image you can imagine with
perfect fidelity in a matter of seconds using just words, though, honestly, is that a parlor trick?
Anyway, one of the biggest examples of how we could go beyond that is in the field of medicine.
One of the canonical examples of a potential AI future utopia is this.
Ten years from now, imagine you go to a doctor's office, and in the examination room there is
video cameras, there are microphones recording everything you say, everything the doctor says,
taking note of all your symptoms, how your cough sounds, all of your interactions, all your scans,
all your x-rays, all your descriptions of your symptoms, they're all being recorded and fed into
an AI system that has been trained on all of the medical literature in history, going back to
the very beginning of time and including stuff published maybe just within the last hour.
Such an AI system would potentially be able to make a diagnosis with 99% or greater accuracy.
And this would be better than the current system where, you know, it's a roll of the dice.
Maybe your doctor is having a bad day or is distracted or misses something in the scan for simple human error.
In essence, the vagaries and fluke inconsistencies of daily life would be eliminated.
You don't roll the dice and get a bad doctor or a bad diagnosis.
and even, think of this example that Mark Andreessen likes to use,
would you prefer to have been seen by an old doctor or a young doctor?
The old doctor has experience, she's seen everything,
but maybe she doesn't keep up on the latest literature like she used to.
The young doctor is not as experienced,
but she's fresh out of school and open to and trained on the latest thing.
So in this scenario of the future, you don't have that Hobson's choice.
The AI makes the best diagnosis, because it's trained again,
on all the medical literature ever in any language, so there's no chance something new hasn't yet
made it to your doctor's consciousness because it's still in Chinese. And crucially, the human doctor
is not made redundant. Their role just changes to something like being a counselor. You can
imagine the doctor saying, okay, Brian, here's what the computer has diagnosed and suggests for
treatment. We have a few options. Let me talk you through them. And then we can come up with a
course of action together. Sounds good. Well, here you go. A step. A step
toward that potential future. Microsoft's speech recognition subsidiary nuance has debuted Dragon Ambient
Experience Express, which uses GPT4 to generate clinical note drafts after patient visits, quoting CNBC.
DAX Express aims to help reduce clinicians' administrative burdens by automatically generating a draft
of a clinical note within seconds after a patient's visit. The technology is powered by a combination of
ambient AI, which forms insights from unstructured data like conversations, and OpenAI's newest model
GPT4. Diana Knoll, the executive VP of Nuisance's healthcare division, told CNBC that the company wants
to see physicians, quote, get back to the joy of medicine so they can take care of more patients.
Our ultimate goal is to reduce this cognitive burden to reduce the amount of time that they
actually have to spend on these administrative tasks, she said.
Microsoft acquired Nuisance for around $16 billion in 2021. The company derives
revenue by selling tools for recognizing and transcribing speech during doctor office visits,
customer service calls, and voicemails. Dax Express complements other existing services that
nuance already has on the market. Null said the technology will be enabled through Nuance's
Dragon Medical One Speech Recognition application, which is used by more than 550,000 physicians.
Dragon Medical One is a cloud-based workflow assistant that physicians can operate using their voices,
allowing them to navigate clinical systems and access patient information quickly.
clinical notes generated by Dax Express will appear in the Dragon Medical One desktop.
Dax Express also builds on the original Dax application that Nuance launched in 2020.
Dax converts verbal patient visits into clinical notes, and it sends them through a human review
process to ensure they are accurate and high quality. The notes appear in the medical record
within four hours after the appointment. Dax Express, in contrast, generates clinical notes within
seconds so that physicians can review automated summaries of their patient visits immediately.
that physicians, clinicians are going to want a combination of all these because every specialty
is different, every patient encounter is different. And you want to have efficient tools for all of
these various types of visits, Noel said, end quote. So today was hopefully the last listener
ad for a while. And I only say hopefully because that would mean our ad inventory has fully
come back and remains healthy. But I don't want to sound like I'm ungrateful. So I want to thank
again those who stepped in and stepped up to help keep the lights on when I needed it. I've been
trying to give some of those folks extra ad impressions again over these last few days when possible
as a thank you, but also patronize the listener ads when you hear them so you can thank them as well.
Talk to you tomorrow.
