Tech Brew Ride Home - Wed. 11/13 – Amazon Haul
Episode Date: November 13, 2024The guessing game on who the new tech regulators will be for the new administration. What is Amazon Haul? Why has Bluesky suddenly seen a flurry of activity? More on Apple’s smarthome ambitions. And... more on why AI seems to be hitting a wall lately? Here’s what you missed today in the world of Tech. Sponsors: Take your personal data back with Incogni! Use code RIDEHOME at the link below and get 60% off an annual plan: Incogni.com/ridehome Links: Donald Trump’s potential antitrust enforcers may keep Big Tech in their sights (Financial Times) Amazon debuts discount store with everything under $20 to take on Temu and Shein (CNBC) With Surge in New Users, Bluesky Emerges as X Alternative (NYTimes) Apple’s Next Device Is an AI Wall Tablet for Home Control, Siri and Video Calls (Bloomberg) OpenAI, Google and Anthropic Are Struggling to Build More Advanced AI (Bloomberg) Bluesky Techmeme Starter Pack 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 Tech meme right home for Wednesday, November 13th, 2024. I'm Brian McCullough.
Today, the guessing game on who the new tech regulators will be for the new administration.
What is Amazon Hall? Why has blue skies suddenly seen a flurry of activity?
More on Apple's smart home ambitions and more on why AI seems to be hitting a wall lately.
Here's what you miss today in the world of tech.
Sources tell the FT that Gail Slater, a top aide to J.D. Vance and Mark Meeter,
A former enforcer at the DOJ's antitrust unit and the FTC are the frontrunners to lead the FTC.
These are the new regulators, perhaps, quote.
Some on Wall Street have expressed worries that Slater and Meeter may be partial to the tough enforcement stance
spearheaded by progressive officials appointed by U.S. President Joe Biden,
FTC Chair Lena Kahn and Jonathan Cantor, head of the DOJ's antitrust division,
which has been embraced by a new generation of Maga Populous with Vance at its vanguard.
either Slater or Meador are likely to disappoint dealmakers hoping for a radical change from the Biden administration,
which has cracked down on anti-competitive conduct across the economy, and taken on big tech giants such as Google, Apple, and Amazon in a bid to reverse what it says have been years of lax enforcement.
Vance, who has pledged to prioritize the working man over Wall Street, could seek to preserve parts of Khan's legacy, despite criticism from dealmakers who have labeled her anti-business, said people following the selection process.
The vice president elect has praised Khan for, quote, doing a pretty good job and said Google should be broken up.
For the DOJ, Trump's transition team might opt for a more traditional antitrust approach that predated Khan and canter,
these people said in an effort to balance opposing wings of the Republican Party over dealmaking.
The leading contenders to head the DOJ's antitrust division are Alex Okuliar, William Renner, and Barry Nigro,
three antitrust lawyers who previously worked with McCann Del Haram when he headed the unit during Trump's first administration.
Under Del Rahim, the DOJ brought high-profile cases against Google, AT&T, and T-Mobile,
and was seen as operating within the more establishment antitrust view that company's growth
can be tolerated so long as consumers are not harmed, a paradigm rejected by Conn and Cantor.
Trump's transition team is consulting Del Rahim to choose contenders for top antitrust jobs,
according to a person close to the transition team.
Slater and Vance are also playing crucial roles in the decision-making.
Slater, who previously worked at the FTC for a decade, was, quote, well-positioned to get whatever
she wants, said someone familiar with the matter. An Oxford-educated lawyer, she was an advisor to Julie
Brill, a former FTC commissioner appointed by Barack Obama, the former Democratic president, end quote.
Amazon has launched Amazon Hall, a storefront on its app that promises crazy low prices
on a range of items to compete with the likes of Timu and Sheehan, quoting CNBC. Shoppers can buy
$1 eyelash curlers and oven gloves or a $3 nail dryer. The company is offering free shipping on orders
over $25 or a $399 shipping fee on orders below that threshold. Amazon is betting shoppers will wait
longer for products in exchange for rock bottom prices. The company noted that most purchases made
in Amazon Hall will be delivered in under two weeks, quote, although shipping times may vary
and are dependent on a customer's delivery location. That's a shift for Amazon, which has partly
cemented its dominance in e-commerce by offering faster delivery speeds than its competitors.
The company upended the online shopping world when it first offered free two-day delivery,
and it's been speeding up delivery times since then.
Amazon now offers same or next day delivery,
and in some parts of the U.S.
it promises delivery within a few hours of an order being placed.
With Amazon Hall, the company is responding to the rise of Temu, Sheehan, and TikTok Shop,
which all have ties to China, the world's second largest economy.
The platforms have rapidly gained popularity in the U.S. over the past few years
by hooking deal-hungry shoppers with their low prices on clothing, makeup, home goods, and other items.
Amazon is able to offer low prices through its Hall storefront
by importing each item directly from manufacturers in China, similar to how Temu and Sheehan's
business models operate. To manage costs, the company says it won't accept returns on items that are
$3 or less, end quote. So I once told you that I would help monitor whether or not there was a
Twitter diaspora, at least for the purposes of this show, i.e. has tech chatter moved to other places?
Well, I got to tell you, there's been a ton of activity on Blue Sky of late. And indeed, the New York Times
says that Blue Sky has gained more than 1 million users in the last week since the U.S.
election, mostly in the U.S., the U.K. and Canada, now has 14.7 million users in total.
Now, what I'm going to quote you from is looking at this through a political lens,
but let me get back to the tech community angle after I'm done reading this.
Quote, we're seeing increased activity levels across all different forms of engagement.
A spokesperson said in an email, Blue Sky, which began in 2019, as a project
by the former Twitter founder Jack Dorsey is among several challengers that have each experienced
burst of momentum since Mr. Musk's acquisition of Twitter in 2022 and subsequent rebranding of it.
The accounts created on Blue Sky this week, many of which are left-leaning, shared cat videos
alongside their hopes that the platform might offer a reprieve from the misinformation and
hateful speech that have swirled on X since Mr. Musk's takeover.
New or freshly active users on the platform include celebrities, the rapper Flavor Flav,
the author John Green, Democratic political,
figures, Representative Alexandria Ocasio-Cortez of New York, Chastin Buttigieg, husband of Pete Buttigieg,
the Transportation Secretary, and media personalities, Medi Hassan, and Molly Zhang Fast.
Hello, less hateful world, Mark Cuban, the billionaire, and Kamala Harris surrogate posted
on Tuesday. Blue Sky gained its first wave of high-profile users last spring and switched
from being invitation only to open to the public in February. It now has 14.7 million users,
the company said, that is still far fewer than threads, Meta's competitor to X,
which this month reported that it had reached 275 million monthly active users.
A met a spokesperson declined to share whether that number had changed post-election.
There has been a, quote, feeling of momentum on blue skies since the election, said Shannon C. McGregor,
an associate professor at the Hussman School of Journalism and Media at the University of North Carolina,
Chapel Hill.
Many of the people leaving acts are disappointed by the outcome of the election, she said,
and now associate the platform with Mr. Musk's support of Mr. Trump.
Others are frustrated by the type of content that seems to be rising to the top of their feeds.
Blue Sky users are doing their best to welcome the flurry of new arrivals.
Many are posting welcome messages, introductions, or starter packs,
list of accounts that might appeal to someone based on their locations or interests, end quote.
Yes, about that last bit.
From what I've been seeing, the tech community is not so much fleeing X for political reasons,
as much as they are fleeing threads for product reasons.
Folks have been complaining for months that the way the Threads algorithm works,
folks can't get as much reach or discussion going as they used to on Twitter.
Folks are saying that Threads is focusing on creating an anodyne community that doesn't lend itself to vigorous debate.
And people are feeling like Blue Sky is offering something closer to what Twitter used to be.
I am at Brian MC over there, not MCC.
For some reason, I got Brian MC when I signed up.
Oh, well, follow me over there.
I might start posting there more soon, but also check the bottom link in the show notes.
One of the tech meme editors put together a starter pack of folks to follow the folks that we follow to monitor the tech chatter.
If you use that starter pack, you should be following most of the tech folks that I'm following.
A couple of segments here that are follow-ons to stuff we've been talking about.
First up, Mark German couldn't let yesterday's speculation from Ming Chi Quo about Apple going heavily into the smart home market go unremarked.
German says Apple is making a six-inch wall-mounted display to control appliances,
video conference, and use AI to navigate apps, and may unveil it as early as March of next year.
Quoting Bloomberg, the company is gearing up to announce the device as early as March
and will position it as a command center for the home, according to people with knowledge of the effort.
The product code named J490 will also spotlight the new Apple Intelligence AI platform
said the people who asked not to be identified because the work is confidential.
Chief Executive Officer Tim Cook is betting that the product can make Apple a force in the smart home segment,
where the company has trailed behind Alphabet and Amazon in recent years.
He has made the device a priority for the company's engineering and design departments
and is pushing to get it to market after more than three years of development.
The device has a roughly 6-inch screen and looks like a square iPad.
It's about the size of two iPhones side by side with a thick edge around the display.
There's also a camera at the top front, a rechargeable built-in battery, and internal speakers.
Apple plans to offer it in silver and black options.
The product has a touch interface that looks like a blend of the Apple Watch operating system
and the iPhones recently launched standby mode,
but the company expects most people to use their voice to interact with the device,
relying on the Siri Digital Assistant and Apple Intelligence.
The hardware was designed around App Intense,
a system that lets AI precisely control applications and tasks,
which is set to debut in the coming months.
The product will be marketed as a way to control home appliances,
chat with Siri, and hold intercom sessions via apps,
Apple's FaceTime software. It will also be loaded with Apple apps, including ones for web browsing,
listening to news updates, and playing music. Users will be able to access their notes and calendar
information, and the device can turn into a slideshow display for their photos. A first for Apple,
the device will compete with Amazon's Echo Show and Echo Hub Smart displays, as well as Google's
Nest Hub. It's also reminiscent of Meta's portal, a failed video conferencing device from the
social media giant. Apple is already planning a more expansive follow.
follow-up version with a robotic limb that can move the screen around. Apple plans to market that
technology as a home companion with an AI personality. The higher-end product could be priced
at as much as $1,000 depending on the components it uses, the people said. The display-only device
will be far less than that, approaching the cost of competitors' products. The Echo Show 8 is
priced at $150, while the Echo Hub is $180. The Nest Hub max costs $230. Apple has designed different
attachments for the device, including ones that affixed the screens onto walls like a classic
home security panel. There will be bases with additional speakers that can be placed in the kitchen
on a nightstand or on a desk. Apple imagines the FaceTime feature being used while cooking or for
video conferencing during work meetings. A person familiar with its development said the product is
designed to bring Siri and Apple Intelligence to life in a way that hasn't happened before, end quote.
And finally, more on AI hitting the proverbial wall, if you will. Quoting Bloomberg again,
Open AI was on the cusp of a milestone. The startup finished an initial round of training in September
for a massive new artificial intelligence model that it hoped would significantly surpass prior versions of the technology behind ChatGPT
and move closer to its goal of powerful AI that outperforms humans. But the model known internally as Orion did not hit the company's desired performance,
according to two people familiar with the matter who spoke on condition of anonymity to discuss company matters.
As of late summer, for example, Orion fell short when trying to answer coding questions that it hadn't been trained on.
The people said, overall, Orion is so far not considered to be as big a step up from OpenAI's existing models as GPT4 was from GPT3.5,
the system that originally powered the company's flagship chatbot, the people said.
OpenAI isn't alone in hitting stumbling blocks recently.
After years of pushing out increasingly sophisticated AI products at a breakneck pace,
Three of the leading AI companies are now seeing diminishing returns from their costly efforts to build newer models.
At Alphabet's Google, an upcoming iteration of its Gemini software is not living up to internal expectations, according to three people with knowledge of the matter.
Anthropic, meanwhile, has seen the timetable slip for the release of its long-awed model called 3.5 Opus.
The companies are facing several challenges.
It's become increasingly difficult to find new untapped sources of high-quality human-made training data that can be used to build more
advanced AI systems. Orion's unsatisfactory coding performance was due in part to the lack of
sufficient coding data to train on, two people said. At the same time, even modest improvements
may not be enough to justify the tremendous costs associated with building and operating
new models or to live up to the expectations that come with branding a product as a major upgrade.
There is plenty of potential to make these models better. Open AI has been putting Orion through
a month's long process, often referred to as post-training, according to one of the people,
That procedure, which is routine before a company releases new AI software publicly, includes incorporating human feedback to improve responses and refining the tone for how the model should interact with users, among other things.
But Orion is still not at the level OpenAI would want in order to release it to users, and the company is unlikely to roll out the system until early next year, one person said.
These issues challenge the gospel that has taken hold in Silicon Valley in recent years, particularly since OpenAI released ChatGPT two years ago.
Much of the tech industry has bet on so-called scaling laws that say more computing power, data, and larger models will inevitably pave the way for greater leaps forward in the power of AI.
The recent setbacks also raise doubts about the heavy investment in AI and the feasibility of reaching an overarching goal.
These companies are aggressively pursuing artificial general intelligence.
The term typically refers to hypothetical AI systems that would match or exceed humans on many intellectual tasks.
The chief executives of OpenAI and Anthropic have recently said,
AGI may only be several years away. The AGI bubble is bursting a little bit, said Margaret Mitchell,
chief ethics scientist at AI startup hugging face. It's become clear, she said, that, quote,
different training approaches may be needed to make AI models work really well on a variety of tasks,
and idea a number of experts in artificial intelligence echoed to Bloomberg News.
Still, AI companies continue to pursue a more is better playbook. In their quest to build products
that approach the level of human intelligence, tech firms are increasing the amount of
computing-powered data and time they use to train new models and driving up costs in the process.
Amadai of Anthropic has said companies will spend $100 million to train a bleeding edge model this year,
and that amount will hit $100 billion in the coming years.
This conundrum has come into focus in recent months inside Silicon Valley.
In March, Anthropic released a set of three new models and said,
the most powerful option called Claude Opus outperformed OpenAIs, GPT4, and Google's Gemini,
systems on key benchmarks such as graduate level reasoning and coding. Over the next few months,
Anthropic pushed out updates to the other two cloud models, but not Opus. That was the one everyone
was excited about, said Simon Willison, an independent AI researcher. By October, Willison and other
industry watchers noticed that wording related to 3.5 opus, including an indication that it would
arrive later this year and was coming soon, was removed from some pages on the company's website.
Similar to its competitors, Anthropic has been facing challenges behind the scenes to develop
3.5 opus, according to two people familiar with the matter. After training it, Anthropic found
3.5 opus performed better on evaluations than the older version, but not by as much as it should,
given the size of the model and how costly it was to build and run, one of the people said,
end quote. As mentioned, the Rad History Podcast, this week is out. It's about the history of
the Rubik's Cube, which this month turned 50 years old. Give that a listen. Enjoy. Talk to you
tomorrow.
