Tech Brew Ride Home - Wed. 12/13 – Big Tesla Recall
Episode Date: December 13, 2023Apple has a new tool to help you if your iPhone gets stolen. There’s a huge Tesla recall related to Autopilot. More AI tools from Google. How is X doing, you know, financially speaking? And Netflix�...��s first ever What We Watched report. Sponsors: Miro.com/podcast Mindbloom.com/techmeme100 and code techmeme100 Links: Apple Makes Security Changes to Protect Users From iPhone Thefts (WSJ) Google Offers Enhanced Generative AI Features for Cloud Customers (Bloomberg) Tesla Recalls 2 Million Cars to Fix Autopilot Safety Flaws (Bloomberg) FCC issues final denial of $885M Starlink subsidy (TechCrunch) Musk’s X 2023 Ad Sales Projected to Slump to About $2.5 Billion (Bloomberg) EU Deal to Force Uber, Deliveroo Treat Some Drivers as Employees (Bloomberg) OpenAI to Pay Politico Parent Axel Springer for Using Its Content (WSJ) Netflix reveals how many hours we spent watching The Night Agent and Queen Charlotte (The Verge) Follow the podcast on TikTok: @techmemeridehomepodcast 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 Wednesday, December 13th, 2023. I'm Brian McCalla today. Apple has a new tool to help you if your iPhone gets stolen. There's a huge Tesla recall related to autopilot. More AI tools from Google. How is X doing? You know, financially speaking. And Netflix's first ever what we watched report. Here's what you miss today in the world of tech. Apple has released a new security setting called stolen device protection. If you enable it, the feature limits actions like Apple ID password.
resets when your device is not in a familiar location. This comes after a whole slew of iPhone
thievery that has been happening recently as described by Joanna Stern in the journal. Quote,
the journal reported on a nationwide spate of thefts where criminals used iPhone owners' passcodes
to change their Apple accounts, access, saved passwords, steal money, and lock them out of
their ICloud stored photos and videos. Thieves in New York, Chicago, New Orleans, Minneapolis, and other
cities watch iPhone users tap in their pass codes before stealing the target's devices. These thefts
resulted in losses far beyond phones, as the journal's reporting showed, because Apple's security
settings gave victims few ways of preventing harm once their pass codes fell into the wrong hands.
We have heard from hundreds of people over the past year whose iPhones and digital lives were
stolen. The new stolen device protection setting designed to defend against such attacks is being
released to beta testers. Apple is planning to include the setting in a coming software
update. Still, users must turn it on, and it won't cover all threats to your personal and financial
information on an iPhone. Your passcode, that short string of numbers that grants access to an iPhone,
has powerful reach. With this number, typically four or six digits, thieves can access a lot of
your data and make sweeping changes to your accounts. And when face ID or touch ID fails,
the passcode serves as a fallback. If you enable the new stolen device protection, your iPhone
will restrict certain settings when you are away from a location familiar to the iPhone,
such as your home or work, end quote.
More AI tool announcements, spelled M-O-A-R. I agree that all of these announces are getting
sort of mind-numbing, so let's just sort of hit the highlights here. Google announced Gemini
Pro for Enterprises, letting developers build apps available for free at launch with limits,
and added Pro to AI Studio and Vertex AI, quoting Bloomberg. The company on Wednesday unveiled
Gemini Pro for enterprises allowing developers to build applications using Google's latest AI model,
which was announced last week. Google Cloud clients can use Gemini to create apps such as
AI-powered chatbots, easy-to-quiry inventory databases, and marketing presentations. The company also
emphasized that Gemini Pro will be free at launch for cloud customers with some limits. Ultimately,
Google said, it plans to ensure that its cloud AI offering will be, quote, competitively priced.
The company said Gemini Pro's text-based capabilities are four times less expensive for input and two
times less expensive for output than the last iteration of its AI model Palm 2, which was released
in June. Gemini Ultra, the company's largest and most capable model for highly complex tasks,
will be available to select cloud customers and partners for early experimentation before its
release to the general public next year, the company said. Google announced that Gemini Pro
would be incorporated into two key cloud products, Google AI Studio and Vertex AI. Google AI Studio,
a free web-based developer tool, is what Google calls the fastest way to build with
Gemini. The tool allows clients to use the Gemini API for developing apps. Vertex AI, meanwhile,
provides developers and cloud clients with more customization. Businesses will be able to customize
Gemini using their company's own data and build Gemini-powered search tools and chatbots,
among other applications. Google also announced upgraded versions of previously released AI
models, MGen 2. Google's text-to-image technology will have improved photorealism,
text rendering, and logo-generation capabilities, the company said. Google is unveiling
MedLM as well, a family of models fine-tuned for the healthcare industry, building on the work
the company has done on Med Palm 2, Google's AI model that has been trained on specialized medical
knowledge, end quote, something, something, AI varietals. Tesla is planning to update autopilot in more
than 2 million vehicles after the NHTSA issued a recall and called Tesla's safety measures inadequate
and open to misuse. This comes after a quite lengthy probe.
by the way, quoting Bloomberg. The move is the result of a year's long defect investigation
that the National Highway Traffic Safety Administration will keep open to monitor the efficacy
of Tesla's fixes. An NHTSA spokesperson said the probe found that Tesla's means for keeping
drivers engaged were inadequate. Automated technology holds great promise for improving safety,
but only when it is deployed responsibly, NHTSA said Wednesday. Today's action is an example
of improving automated systems by prioritizing safety, end quote.
In its recent recall report, Tesla has announced plans to implement an over-the-air software update
set to roll out around December 12th. This update is aimed at enhancing control features and safety alerts
within their automated driving systems. This recall marks the second such action taken by
Tesla this year as their automated driving technology faces mounting scrutiny due to a series
of accidents, some of which have resulted in fatalities. The NHTSA had previously investigated
Tesla's autopilot system after a fatal 2016 crash, but cleared it early in the following year.
However, the NHTSA is now conducting two ongoing defect probes initiated in August 2021 and February
2022, both triggered by incidents involving Tesla's colliding with first responder vehicles and
sudden braking on highways. The pace of these investigations has increased under the Biden
administration with over 50 special crash investigations involving Tesla vehicles
potentially linked to autopilot. Beyond the NHTSA, other regulatory bodies are
are scrutinizing Tesla's driving systems. In January, Tesla revealed that it had been requested to
provide documents to the Justice Department concerning autopilot and full self-driving beta.
Additionally, Bloomberg reported that the Securities and Exchange Commission was conducting an
investigation into Elon must involvement in shaping Tesla's claims about self-driving technology.
Speaking of Elon's various businesses, the Federal Communications Commission has issued a final
denial of Starlink's request for $885 million to expand
internet coverage to rural areas, saying Starlink, quote, failed to demonstrate an ability to
offer the service. Quoting TechCrunch. As previously reported, the money in question was part of the
Rural Digital Opportunity Fund, a multi-billion dollar program to subsidize the rollout of internet service
in places where private companies have previously decided it's too expensive or distant to do so.
The $885 million was first set aside for Starlink in 2020, corresponding to the company's bid on how
much connectivity it could provide, at what cost, and to which regions. The FCC explained that this
first application was a high-level, short one, and that those qualifying for that would receive
closer scrutiny. For instance, one organization assigned over a billion dollars in funds turned out
to be a regional operation that couldn't possibly expand the way it hoped to, end quote.
Last summer, the FCC's review of Starlink satellite internet proposal found it to be a, quote,
still developing technology that required users to purchase a $600 dish, which posed a financial
hurdle for many potential customers, especially those with limited resources.
The FCC considered not allowing orbital communications companies to apply, but ultimately
decided to let them compete based on their merits. The FCC identified several, quote,
financial and technical deficiencies in Starlink's proposal and operations. Starlink
contested this decision, asserting that it had been subjected to, and quote, inappropriately
onerous standard. While certain details are redacted, Starlink argued that despite initial tests
showing declining speeds and other issues, it had plans to expand its satellite network by launching more satellites.
The promise of SpaceX's super-heavy launch vehicle, Starship, was presented as evidence of those expansion plans.
Two FCC commissioners, Brendan Carr and Nathan Symington, disagreed with the decision to reject Starlink's proposal.
Symington highlighted that many of the recipients of the World Digital Opportunity Fund did not provide any service at all,
whereas Starlink was already serving half a million subscribers, including areas with limited broadband options.
He criticized the Bureau's reasoning for its decision, considering the launch problems as minor issues.
Hey, since we've been doing a sort of whip around Elon's various businesses, it's been a while since we checked in with X to see how it's going over there.
Is Elon turning things around? Well, sources have told Bloomberg that X generated just over $600 million in ad revenue in each of 2023's first three quarters.
That's down from more than a billion dollars per quarter back in 2022, thereby putting X,
X on track to make around, let's say, $2.5 billion in ad revenue this year, quoting Kurt Wagner
in Bloomberg. The company also makes money from its subscription service, X premium, and from data
licensing agreements. External estimates peg the subscription business at less than $120 million annually.
In 2021, X then Twitter, generated $572 million in revenue from data licensing deals.
X was not profitable when Musk took over, but the company reported more than $5 billion in revenue the
year before Musk acquired the company. In early 2021, Twitter executives led by then-CEO
Jack Dorsey set a public goal to reach $7.5 billion in revenue by the end of 2023, end
quote. Not going to make that. So from more than $5 billion a year in revenue before Elon to
maybe, let's call it $3.4 billion-ish in revenue this year, when you factor in subscriptions
and whatever data licensing deals are left, because remember Elon blew a lot of those up,
I don't know. Doesn't seem like you could call this a term.
around quite yet. More European regulatory stuff, the EU Council and EU Parliament have agreed a deal
to reclassify millions of ride-hailing and food delivery app workers as staff, possibly costing the
industry billions. Quoting Bloomberg, the provisional deal could require platforms to give full status
to the estimated 5.5 million workers who meet at least two out of five conditions that indicate
their relationship with the apps fits that of an employee, rather than someone who's self-employed,
the EU said in a statement on Wednesday. Those conditions,
include the platform setting upper limits on worker pay, monitoring performance, controlling the
distribution of tasks and working hours, and making rules about staff appearance or conduct.
The EU's deal would give eligible individuals' rights typically enjoyed by employees,
including minimum wage, paid leave, and unemployment benefits. Still, there are fears that
stricter employment rules will push the delivery platforms to cut back. A similar law passed in
Spain two years ago, prompting Deliveroo to pull out of the country and other food delivery
afts to reduce their operations there. Wednesday's provisional agreement must still be endorsed and adopted
by the Council of the EU and Parliament. Member states will have two years to incorporate the rules
the EU said, end quote. In this year of AI ascendancy, one of the things we maybe haven't talked
enough about is the whole question of, will AI put the final nail in media's coffin, because now
the bots can write all the news? Or, if data is the new, new oil, and LLM's
need content to train on, could this be a rich new vein of hidden revenue that could maybe save
media? So expect a lot more media companies to dip their toes in the water sort of like this.
Axle Springer has signed a multi-year licensing deal with OpenAI, which will pay Axel Springer
to use content from Politico, Business Insider, and others to train its AI models, quoting the journal.
The companies declined to disclose financial terms, but the pact is expected to generate substantial
revenue for Axel Springer. When ChatGPT uses information from Axel Springer publications to answer
a user's query, it will include links to the original sources of the information below the answer,
the company said in a joint statement. The new format, which would generate an answer in the form of a
summary, is set to debut in the coming months. The aim is to ensure that those websites get
credit, compensation, and web traffic. OpenAI will be able to display answers and train its
technology using Axel Springer's archival and current material. Axel Springer will also be able to use
OpenAI's technology to improve its own products, the publisher said. We want to explore the opportunities
of AI-empowered journalism to bring quality societal relevance and the business model of journalism
to the next level. Axel Springer chief executive Mattias Duffner said in a statement,
the deal isn't exclusive. That means Axel Springer can generate revenue from deals with
other generative AI companies that want to use its content. In its negotiations with publishers,
OpenAI has discussed using article word count as one of the measures to determine compensation
for use of their content, according to people familiar with the matter.
The Associated Press recently announced its own licensing deal with OpenAI.
OpenAI is licensing part of AP's text archive for training,
but the deal doesn't include the use of AP content in chat GPT,
according to a person familiar with the matter, end quote.
Finally today, maybe they just didn't want Spotify to have all the year-end fun.
Netflix has published its first What We Watched report,
detailing the most watched content from January to June of this year.
Quoting The Verge.
Netflix is going to start publishing a new report twice a year that details the most popular
shows and movies on the platform.
The first report released today details the most watched content from January to June,
and it's perhaps the best look yet at how much people are actually watching Netflix's
gargantuan library of titles.
What We Watched, a Netflix engagement report will track three metrics, hours watched,
whether a show is available globally and a show's release date.
In this first report, the first season of The Night Agent topped the list with more than
812 million hours viewed, followed by Ginny and George's second season, 665.1 million hours viewed,
the glory's first season, 622.8 million hours viewed, Wednesday's first season, 5707 million
hours viewed, and Queen Charlotte, a Bridgerton story, 503 million hours viewed. Those are only the top
five. The full list contains more than 18,000 titles, and content is included if it has been
watched for more than 50,000 hours. Netflix already offers its weekly top 10 lists, but this new report
is a much wider view of Netflix's viewership data. Streaming data was a key point of contention
during the recent Writers Guild of America Strike, and the new contract requires that Hollywood
studios must share streaming data with the guild, end quote. Yeah, but why doesn't Netflix do
something like Spotify wrapped? How many hours you spent on certain shows or even certain genres
in a year? Though now that I think about it, maybe we wouldn't care about that. Like,
seeing what you listen to in a given year gives you kind of a sense of your subconscious in a way,
and they're the sense of curation and your musical taste.
It somehow feels like music is more defining of who you are.
Maybe TV just isn't like that.
All right, here's the big show news announcement that I've been promising.
When the whole Sam Altman Open AI drama happened,
I had a real crisis of confidence because, you know, I do this show for a living,
but I actually care about it.
I care about delivering news to you.
That's why you don't just get a show that ends exactly.
exactly at 15 minutes every day.
If something happens that I think is cool, I want you to know about it, so I'll tell you about it.
So that whole week was really hard for me.
This was the biggest tech story of the year, and I felt like I was letting you down because I wasn't covering it regularly, or more comprehensively, I guess.
The most recent bonus episode was my attempt to make up for that.
But it got me thinking, could I build a mechanism to deliver breaking news to you without interrupting the usual cadence of the podcast for?
feed. Like, I don't want to be doing little 30-second emergency episodes all the time, but sometimes
something happens, and I do want to tell you about it right away. So, the more I thought about this,
the more an idea sparked. 30-second emergency updates. What is 30-second or so roughly TikToks, right?
So what if I had a TikTok account that you could follow to give you breaking news summaries in like
30 seconds? Since Thanksgiving, I've been noodling away at how to do that, and here's what I've come up with.
The TikTok account is TechMeme Ride Home Podcast, all one word, no spaces, no underscores,
TechMeme Ride Home Podcast.
Go look that up and follow that account.
And what I think I'll do is, you know, just release a few TikToks every day.
As I'm writing a segment, I could do a quick 30-second post showing you the article I'm thinking
of going with, and also the tweets and chatter that sent me there.
So in real time, I could show you what I'm working on that day, or I could do like two to three quick
TikTok's in the morning when I first sit down to work on the show and give you a heads up on what
you'll probably be hearing about later. And even at night, if something major breaks, I can
quickly shoot you a TikTok so even though you won't have me give you the full rundown and context
for another 18 hours or so, you will at least be in the know that something happened right
away. That's always been on my mind with this show. I'm only one guy, so obviously I can't
deliver news in real time all the time. A lot of times I have to tell you about stuff the day
after or even more. And I know you come to me for analysis and contextualization and summary,
just as much as breaking news. But, well, think of this as an effort in a sort of supplementary way,
sort of to fulfill that real-time news value proposition. So if you use TikTok, follow the account,
TechMeme Ride Home podcast, again, all one word, because there is a similar account with underscores
that I set up to do some experiments with last year. Ignore that one. I'll put a link to the one you
want to follow in the bottom of the show notes. You'll know that you've hit the right account
if the couple most recent videos are me as the Grinch, Grinch filter turned on. And be sure to
tap the heart on the videos for the next few days and try to watch them all the way through
because apparently the TikTok Algo really waits view completion heavily. Help me grow the
account quickly by, you know, viewing all the videos all the way through. If this proves useful
to people, maybe I'll set up other accounts as well, Instagram, Reels, YouTube shorts,
I don't know, Snapchat threads.
Wherever people are, I could post these videos because once I've made the video, you know, I have it.
I can just put it out everywhere.
But for now, let's focus on that TikTok account, please.
Tech meme, ride home podcast, all one word.
Talk to you tomorrow.
