Category: TECH

Following Elon Musk’s xAI’s move to open source its Grok large language model earlier in March, the X owner on Tuesday said that the company formerly known as Twitter will soon offer the Grok chatbot to more paying subscribers. In a post on X, Musk announced Grok will become available to Premium subscribers this week, not just those on the higher-end tier, Premium+, as before.

The move could signal a desire to compete more directly with other popular chatbots, like OpenAI’s ChatGPT or Anthropic’s Claude. But it could also be an indication that X is trying to bump up its subscriber figures. The news arrives at a time when data indicates that fewer people are using the X platform, and it’s struggling to retain those who are. According to recent data from Sensor Tower, reported by NBC News, X usage in the U.S. was down 18% year-over-year as of February, and down 23% since Musk’s acquisition.

Musk’s war on advertisers may have also hurt the company’s revenue prospects, as Sensor Tower found that 75 out of the top 100 U.S. advertisers on X from October 2022 no longer spent ad budget on the platform.

Offering access to an AI chatbot could potentially prevent X users from fleeing to other platforms — like decentralized platforms Mastodon and Bluesky, or Instagram’s Threads, which rapidly gained traction thanks to Meta’s resources to reach over 130 million monthly users as of the fourth quarter 2023.

Musk didn’t say when Grok would become available to X users, only that it “would be enabled” for all Premium subscribers sometime “later this week.”

X Premium is the company’s mid-tier subscription starting at $8 per month (on the web) or $84 per year. Previously, Grok was only available to Premium+ subscribers, at $16 per month or a hefty $168 per year.

Grok’s chatbot may appeal to Musk’s followers and heavy X users as it will respond to questions about topics that are typically off-limits to other AI chatbots, like conspiracies or more controversial political ideas. It will also answer questions with “a rebellious streak,” as Musk has described it. Most notably, Grok has the ability to access real-time X data — something rivals can’t offer.

Of course, the value of that data under Musk’s reign may be diminishing if X is losing users.

In this edition of Week in Review, we have big news on the latest Apple antitrust lawsuit. Microsoft’s AI ambitions dominated the news as well, so let’s dive right in.…

Did someone forward this to you? Sign up here to receive the Week in Review newsletter in your inboxes.

The U.S. joined international regulators in accusing Apple of using monopolistic tactics to lock in iPhone users. In response, Apple claims the DOJ’s actions could ruin exactly what its users enjoy about its phones and ecosystem.

The DOJ’s lawsuit draws a connection between Apple and the 1990s era of Microsoft, though there are some key differences between the two companies and their efforts to retain customers.

All that said, don’t expect much to come out of the lawsuit in the short-term. Experts predict it could take three to five years for it to reach a resolution.

But that’s not all — read on for more updates on Reddit’s IPO, Nvidia’s GTC event and an unexpected change for Glassdoor accounts.

News

Microsoft absorbs Inflection AI’s leads: The co-founders of the high-profile AI startup were scooped up by Microsoft on Tuesday in a deal that positions Mustafa Suleyman as the lead of consumer-facing unit Microsoft AI, and Karén Simonyan as the EVP and CEO of the same group. This all comes after Inflection raised $1.3 billion, with the biggest investor in the startup being, you guessed it, Microsoft.

Nvidia’s GTC event: Remember when their biggest news was tied to ray tracing? In their wide-ranging GTC developer conference, Nvidia had a couple of surprises from CEO Jensen Huang, including a prediction that artificial general intelligence was five years away and news of a new AI platform for humanoid robots called GR00T.

Google’s Gemini AI on iPhones?: The two companies are reportedly in talks that could lead to Google’s AI model being deployed to power several upcoming iOS updates, which leaves the question of whether this will be a stopgap until Apple’s own AI efforts are up to speed, a longstanding partnership, or just one of those negotiations that never reaches the finish line.

Hacking & privacy

Investors undercut by past malware investments: The Biden administration is bringing together an international coalition to fight against commercial spyware, with investors now hopping on board. But one of those investors was previously involved in the very business they’re now fighting, TechCrunch exclusively learned.

Hacking a $5 million tournament “for fun”: Sometimes, the stated justification for a hack that takes the gaming community by storm is as simple as that, plus an intent to highlight for developers the exploit used to grind an Apex Legends tournament to a halt.

Be careful about those Glassdoor reviews: Users are reporting that their real names are being added to their profiles without their consent, and the only solution provided to them is to delete their accounts entirely. Time to double-check that burner you made to bad-mouth that ex-employer!

Funding & IPOs

Reddit’s IPO gets off to a strong start: The stock, which launched at $34 a share, jumped 48% on its first day and ended up at $46 after the market closed Friday.

Astera Labs balloons 72% on day one: The company, which makes connectivity hardware for cloud computing data centers, is benefiting immensely from the AI boom, with revenue growing to 45% to $115.8 million last year and its stock closing at $62.03 on day one.

The Browser Company raises $50 million at $550 million valuation: The startup behind the Arc browser has an ambitious, complicated and at-times controversial take on supplanting PCs with browsers. Amid numerous feature launches and AI updates, the company remains unclear about its path to monetization.

Bonus round

AI is bad at spelling, and image generators are to blame: If you’ve ever seen awful spelling in an AI-generated image, you’re not alone. We dove deep into why, despite its massive potential, AI still has trouble spelling the word “burrito.”

Fisker pauses production: And it’s not because of the EVs — it’s because the company is swiftly running out of money. If they’re not able to raise more capital, they could cease operations altogether.

Why AI can’t be reviewed, but we’ll still try: The fact that systems like ChatGPT or Gemini are impossible to truly review makes it all that much more important to put them to the test.

SpaceX is in the final stages of certifying a second pad for astronaut launches, which should ease launch site congestion and help the company scale the number of humans it sends to space.

SpaceX has performed 13 crewed missions, and all of them launched from Launch Complex 39A at NASA’s Kennedy Space Center. It’s the only pad currently certified for human spaceflight. But the company has long intended to upgrade a second pad — SLC-40 at the neighboring Cape Canaveral Space Force Station — to expand its crew launch capacity.

The company is nearly there. Last fall, SpaceX workers installed a crew access arm to the launch tower, a key piece of infrastructure that allows astronauts access to the crew Dragon spacecraft. The company also installed a new emergency egress system at the pad to allow the crew a quick escape in the case of an anomaly.

The system, which is essentially a long orange slide that stretches from the crew tower to the ground, will help SpaceX “scale to bigger towers and spaceships (think 100 people on Starship),” VP of Launch Kiko Dontchev said in a social media post.

As a next step, SpaceX is going to launch the CRS-30 cargo resupply mission. That mission, which is part of a series of now-routine cargo missions the company performs for NASA, is due to lift-off tomorrow at 4:55 p.m. EDT. As the name of the mission suggests, it is the company’s 30th mission delivering essential materials to astronauts aboard the ISS since 2012.

In a media teleconference in February, SpaceX’s VP of build and flight reliability, Bill Gerstenmaier, said the cargo flight is an incremental step to astronaut launches. Both cargo and crew missions use variants of the company’s Dragon spacecraft.

“We would like to do a cargo flight first if we can. We think CRS-30 is probably the right time to do that,” he said. “The work’s pretty much completed at the pad. [We’ve] got some stuff to do next week, but we’ll be in good shape for CRS-30.”

Arjun Sethi speaks with the confidence of someone who knows more than other people, or else who knows that sounding highly confident can shape perception. Either way, when he tells me over Zoom that “in five years, I’ll have 50% of the world’s private data” at his fingertips, and that it will be “impossible to compete against me,” he says it with all the authority of Warren Buffett dropping stock market truths.

Sethi, a co-founder of the venture firm Tribe Capital, is talking about Termina, a subscription-based AI software platform for “quantitative diligence” that Tribe more recently spun out and which Sethi says can improve results for as many investors as he can sign up. Yet the mere prospect raises questions. Namely, if Termina is so good, why are Sethi and Tribe giving other investment firms a way to better compete? Relatedly, why should other investors trust Termina, which ingests its customers’ data to improve over time?

First, it’s worth learning more about Termina itself, which soft launched last month with two products. At a high level, one of those products is a dashboard that aims to help investors quickly gauge the health of any company by comparing it across the companies in Termina’s initial proprietary dataset, combined with the customers’ own data. The second offering is designed to help investors understand external forces at play, including expected market changes. Think of Termina as “giving you the power of 1,000 associates,” says Sethi.

Sethi declines to disclose early customer names, but he says they include pension funds, sovereign funds, and private equity funds, all of which he collectively characterizes as a “large class of capital deployment folks who want to stay under the radar.”

What he’s happier to discuss are his ambitions for the business, whose biggest starting advantage, says Sethi, is transactional data on roughly 1,500 companies that Termina has fed into more than one large language model (that Sethi declines to name), along with an LLM that he says Termina has itself created to enhance investors’ benchmarking capabilities.

To leverage the tool, Termina’s customers also give the outfit their own raw data. “They might be using it for M&A, so a private equity firm will dump their own data [into it] and the same is true of any other VC firm,” he says. What that software spits out, ostensibly, are comparison datasets of companies that exist “across the board, across all stages,” says Sethi, who calls Termina, which was founded just six months ago – the “true Bloomberg of private markets.”

Naturally, some will scoff at the prospect of Termina handling so much sensitive data, which is every firm’s most precious resource. This may prove doubly true given Termina’s ties to Tribe Capital.

Still, he dismisses such concerns. For one thing, he says, Tribe invests in seed- through Series C-stage companies and looks for venture-like returns. Termina, meanwhile, “is enabling anybody in any asset class to be able to think about how to invest into software companies, regardless of stage, so it’s two very different strategies.”

Sethi – who sold an earlier company called MessageMe to Yahoo, then worked with VC Chamath Palihapitiya at his firm Social Capital before co-founding Tribe – also suggests he has the reputation needed to pull off what he’s attempting.

“Part of it,” Sethi says of persuading other investors to use Termina, “is you just inherently build a lot of trust over a long period of time.”

We’ll see. Certainly, one can see both the advantages and disadvantages of Termina’s close relationship with Tribe, which owns a stake in Termina. (Termina has so far raised slightly more than $10 million, including from Sethi personally.)

A team that has long worked together is one advantage, for example. In addition to Sethi – who stepped down as Tribe’s CEO last December, assuming the role of chairman and chief investment officer instead to focus in part on Termina – the seven-person startup includes others who split their time between Termina and Tribe.

Among these is Alex Chee, who cofounded MessageMe with Sethi, joined him at Social Capital, and subsequently co-founded both Tribe and Termina with him. Chee oversees Termina’s day-to-day operations; he is also still Tribe’s chief product officer. Jake Ellowitz, another Termina co-founder and, for a short time, a Social Capital employee, is Tribe’s CTO.

Conversely, Tribe itself is fairly young and has yet to really prove itself. The 30-person outfit boasts an impressive $1.6 billion in assets under management already, but it was founded just six years ago. It has enjoyed some success from investments in crypto tokens, but its other investments have yet to exit and the trajectory of some of those bets is unclear, highlighting the limits of quantitative analysis. Take Carta, the cap table management company that has come under fire for tactics that some see as questionable, or Bolt, the one-click checkout company that was buzzy until it wasn’t.

It’s also worth noting Termina’s clients aren’t agreeing to share their data with the outfit exclusively, so if someone’s black box is better, Termina could be toast.

Again, Sethi shakes off talk of challenges. “The reason why we exist and why customers work with us is that we have the best data in the world, and we have the best product. We don’t have any specific patents. Everything we do is trade secrets. And our tool is not 10x better than what’s out there; it’s, like, one thousand times better than what people’s existing workflow looks like.”

As for what that means for Tribe, he apparently has bigger ambitions now. “My whole goal is that as one venture firm, I could only do so much. As a company, I can just do that much more.”

Texas residents can no longer access Pornhub — without a VPN. As of Thursday, when people in Texas try to access any of the porn sites owned by Aylo, formerly MindGeek, they’ll instead see a long message in opposition to age verification laws.

Laws requiring age verification on adult entertainment sites have already gone into effect in states including Louisiana, Mississippi, Virginia, Utah, Montana and North Carolina. But Texas’ age-verification law was stalled when a group of porn industry advocates and companies challenged the legislation, leading a U.S. District judge to rule that enforcement could harm free speech protections. Last week, the U.S. Court of Appeals ruled that Texas can indeed enforce this law.

Though these laws are intended to prohibit minors from accessing adult content, advocacy groups like the Electronic Frontier Foundation and think tank R Street have long opposed website age checks because they can threaten data security and anonymity. Rather than instituting age checks, which Pornhub has called “haphazard and dangerous,” the site has simply cut off access in states with such laws.

“Since age verification software requires users to hand over extremely sensitive information, it opens the door for the risk of data breaches,” Pornhub wrote on its blog. “Whether or not your intentions are good, governments have historically struggled to secure this data.”

Pornhub didn’t always take this approach. Louisiana was the first state to institute one of these laws at the start of 2023. At first, Pornhub complied with age verification via the state-run LA Wallet app, which connects to a Louisiana resident’s ID. In other words, you were legally required to verify your government ID in order to watch porn.

Once these laws extended to other states, Pornhub and Aylo’s other sites took a different approach. In some states that require age checks to visit adult websites, you’ll be greeted by a safe-for-work video from porn star Cherie DeVille, where she explains Pornhub’s stance on the data privacy issues at hand.

The adult actors who live in these states can still use Pornhub through their creator accounts, because they’ve already verified their identity as a pre-requisite to uploading content on the site.

These changes in state internet regulations come at a time when access to online platforms is up for debate in Congress. The Kids Online Safety Act and similar bills are picking up steam in Congress after numerous hearings with tech CEOs about children’s internet safety, but the common refrain is that many of these safety initiatives could backfire when applied by bad actors. Meanwhile, the House of Representatives passed a bill this week that could ban TikTok and other “foreign adversary controlled applications.” Now, the TikTok bill will be debated in the Senate, where it could have wider ramifications for government interference in tech.

Cyberattacks, regional conflict, weapons of mass destruction, terrorism, commercial spyware, AI, misinformation, disinformation, deepfakes, and TikTok. These are just some of the top perceived threats that the United States faces, according to the U.S. government’s intelligence agency’s latest global risk assessment.

The unclassified report published Monday — sanitized for public release — gave a frank annual window into the U.S. intelligence community’s collective hive mind about the threats it sees facing the U.S. homeland based on its massive banks of gathered intelligence. Now in an election year, the top U.S. spies increasingly cite emerging technology and cybersecurity as playing a factor in assessing its national security posture.

In an unclassified session with the Senate Intelligence Committee on Monday, the top leaders across the U.S. government’s intelligence agencies — including the FBI, NSA, CIA, and others — testified to lawmakers largely to answer their questions about the current state of global affairs.

Here’s what we learned from the hearing.

At least 74 countries use commercial spyware

In the last few years, the U.S. government turned its attention to the government spyware industry, currently made of companies like NSO Group and Intellexa, and previously Hacking Team and FinFisher. In its annual report, the intelligence community wrote that, “from 2011 to 2023, at least 74 countries contracted with private companies to obtain commercial spyware, which governments are increasingly using to target dissidents and journalists.”

The report does not clarify where the intelligence community got that number from, and the Office of the Director of National Intelligence did not respond to a request for comment asking to clarity

But last year, the Carnegie Endowment for International Peace, a Washington D.C. think-tank, released a report on the global spyware industry that included the same number of countries as well as the same dates as the new intelligence community report. The Carnegie report, written by Steven Feldstein and Brian Kot, referenced data that the two collected, which they said came from sources such as digital rights groups and security researchers that have studied the spyware industry like Citizen Lab, the Electronic Frontier Foundation, and Privacy International, as well as news reports.

It’s important to note that the Carnegie dataset, as the authors explained last year, includes what we refer to as government or commercial spyware, meaning tools to remotely hack and surveil targets remotely, such as those that NSO and Intellexa make. But it also includes digital forensic software used to extract data from phones and computers that are physically in the possession of the authorities. Two of the most well known makers of this type of tools are Cellebrite and Grayshift, both of which are widely used in the United States as well as in other countries.

U.S. says it’s struggling to counter ransomware

The U.S. says ransomware is an ongoing risk to U.S. public services and critical infrastructure because cybercriminals associated with ransomware are “improving their attacks, extorting funds, disrupting critical services, and exposing sensitive data.”

Ransomware has become a global problem, with hacking gangs extorting companies in some cases millions in ransom payments to get their stolen files back. Some cybersecurity experts have called on governments to outright ban ransom payments as necessary to stop hackers profiteering from cybercrime.

But the U.S. has shunned that view and takes a different approach, opting to systematically disrupt, dismantle and sanction some of the worst offenders, who are based in Russia and outside of the reach of U.S. justice.

“Absent cooperative law enforcement from Russia or other countries that provide cyber criminals a safe haven or permissive environment, mitigation efforts will remain limited,” the threat assessment reads. In other words, until Russia — and a few other hostile states — give up their criminals, expect ransomware to continue to be the modern-day snow day.

U.S. warns of growing use of AI in influence operations

The use of generative AI in digital influence operations isn’t new, but the wide availability of AI tools is lowering the bar for malicious actors engaging in online influence operations, like election interference and generating deepfakes.

The rise of detailed and convincing deepfake imagery and video is playing its role in information warfare by deliberately sowing confusion and discord, citing Russia’s use of deepfake imagery against Ukraine on the battlefield.

“Russia’s influence actors have adapted their efforts to better hide their hand, and may use new technologies, such as generative AI, to improve their capabilities and reach into Western audiences,” warned the report.

This was something echoed by NSA cybersecurity director Rob Joyce earlier in January about how foreign hackers are using chatbot tools to generate more convincing phishing emails, but that AI is also useful for digital defense.

The report also noted that China is increasingly experimenting with generative AI, noting that TikTok accounts run by a Chinese military propaganda arm “reportedly targeted candidates from both political parties during the U.S. midterm election cycle in 2022.”

There are no laws limiting U.S. spies from buying Americans’ data

U.S. spy agencies have caught on to a popular practice: Why get a warrant for data when they can just buy it online? Given how much data we share from our phone apps (which many don’t give a second thought), U.S. spy agencies are simply buying up vast troves of Americans’ commercially available location data and internet traffic from the data brokers.

How is that legal? After a brief exchange with the head of the Defense Intelligence Agency — one of the agencies confirmed to have bought access to a database containing Americans’ location data — Sen. Ron Wyden noted that the practice was allowed because there is no constitutional or statutory limit on buying commercially available data.

In other words, U.S. spy agencies can keep buying data on Americans that is readily available for purchase until Congress puts a stop to the practice — even if the root of the problem is that data brokers shouldn’t have our data to begin with.

Bumble has lost a third of its Texas workforce in the months since the state passed the controversial abortion SB 8 (Senate Bill 8), also known as the Texas Heartbeat Act, over a year ago. This new data point was shared by Bumble’s Interim General Counsel, Elizabeth Monteleone, speaking on a panel this afternoon at the SXSW conference in Austin, Texas. The panel focused on the “healthcare crisis in Post-Roe America” and featured women who had both sued and spoken out about the need to have doctors, not politicians, involved in their healthcare decisions.

What’s more, Monteleone noted that Bumble is no longer requiring employees to join the business in its Austin location, even though the dating app maker is headquartered there.

“We are a remote-first company. We’ve supported employees who’ve chosen to move out of state,” Monteleone added.

“We — since SB 8 — have seen a reduction in our Texas workforce by about a third. Those employees are choosing to move elsewhere,” she told the audience at the event. “There are a variety of laws in Texas that I think many people find incompatible with living a healthy life and being their authentic self,” added Monteleone, suggesting that not all the departures may be tied directly to this specific piece of legislation, but possibly to several other Texas laws or proposed laws that don’t sit well with Bumble’s employees.

The dating app maker became the first business to join an amicus brief in support of a lawsuit against the Texas abortion law, Zurawski v. State of Texas, filed by the Center for Reproductive Rights. The suit claims that the law puts the lives of pregnant women in jeopardy because doctors are afraid to offer abortion care for fear of losing their licenses, facing hefty fines, or even prison terms. Women involved in the case are suing the state for being forced to carry out their pregnancies because of the state’s abortion law, despite risks to their health. Some women had to travel out of state to get health care, increasing their health risks. Others sued because had to carry non-viable pregnancies to term. Several Texas doctors signed onto the lawsuit, as well, saying they could no longer properly practice medicine.

While the SXSW panel largely focused on the political aspects of this and other laws, including those that now seek to restrict access to IVF, as well as their personal and emotional toll on women, Bumble’s lawyer additionally pointed to the business impact these sorts of laws have.

“We know that abortion has an impact on individuals, but there’s a profound negative impact on society as a whole and particularly for businesses,” Monteleone said. “We were looking at the lens and what we are putting forth in the brief was talking about the increased cost to us to attract and retain talent in Texas, the increased costs for us to provide health care benefits to our employees…We found that because of our position, having been founded with women at the forefront, from the beginning, and having that voice and that legitimacy, to begin with, we could combine that with this very business-centric argument and help support the case,” she said.

Since Bumble’s signing, businesses from across Texas have also signed onto the amicus brief, including rival Match Group and SXSW, some saying they fear they will have similar challenges in attracting and retaining talent, as well.

“These are considerations of [prospective employees] when they’re thinking about if they’re going to accept a job, if they’re going to stay at a job. That consideration about location is a very real factor,” noted Monteleone.

Since the law’s passing, Bumble has introduced enhanced health care benefits that include covering the costs for individuals who “seek the full spectrum of reproductive rights,” she said, including abortion, IVF, surrogacy, egg freezing, and also gender-affirming care. Other companies have offered similar benefits, at their own expense.

These issues are even more pressing for a company like Bumble which is currently struggling with growth and appealing to a younger audience that seems less interested in dating apps than their older counterparts. The dating app maker posted a weak Q4, with a $32 million net loss and $273.6 million in revenue. It also announced it was letting go of 350 employees after other organizational shifts that saw founder Whitney Wolfe Herd stepping down as CEO and a shake-up in the C-suite, which included the appointment of former Slack CEO Lidiane Jones as its new CEO.

Last month, Amazon announced that it’d launch a new AI-powered chatbot, Rufus, inside the Amazon Shopping app for Android and iOS. After a few weeks’ delay, the company began to roll out Rufus to early testers today — including some of us at TechCrunch — to help find and compare products as well as provide recommendations on what to buy.

So I put it through the ringer, naturally.

Rufus can be summoned in one of two ways on mobile: by swiping up from the bottom of the screen while browsing Amazon’s catalog or by tapping on the search bar, then one of the blue-bubbled suggestions under the new “Ask a question” section. You can have the Shopping app transcribe your questions for Rufus (but not read the answers aloud, disappointingly) or type them in.

The Rufus chat interface is pretty bare bones at the moment. There’s a field for questions… and that’s about it. Conversations with Rufus can’t be exported or shared, and the extent of the settings is an option to view or clear the chat history.

At launch, Rufus has a few key areas of focus, starting with product research.

If you’re interested in buying a specific thing (e.g. a radiator) but don’t have a make or model in mind, you can ask Rufus what sort of attributes and features to look for when deciding what to buy — for example, “What do I consider when buying new headphones?” Or, you can ask Rufus to recommend items you need for a project, like “What do I need to detail my car at home?”

Along these lines, I asked Rufus for general buying advice:

• What are the best smartphones?
• Recommend breakfast cereal.

Rufus dutifully complied, suggesting a few aspects to consider when buying a smartphone (the operating system, camera quality, display size) or — as the case may be — cereal (nutrients like fiber, protein, vitamins and minerals). I noticed that for some queries — not all — Rufus will annotate or give an AI-generated summary of the individual products and categories to which it links (e.g. “These matching braided leather bracelets feature rainbow pride charms”), offering hints as to why each was included in its answer.

Rufus recommends cereal.

Curious to see how Rufus would do with more narrow searches, I asked:

• What are the best laptops for teenagers?
• What are the best Valentine’s Day gifts for gay couples?
• What are the best cheap leather jackets for men?
• Recommend books for men.
• Recommend books for women.
• What is the best-reviewed cheap vacuum?

Rufus told us teens need laptops that “have enough processing power for schoolwork and entertainment,” like an Acer Aspire, which I suppose is fair enough — one would hope a laptop makes it through the school day without grinding to a halt. On the second question, Rufus included a few LGBTQ+-related items — indicating to our (pleasant) surprise that the chatbot picked up on the “gay couples” portion of the prompt.

Rufus gives Valentine’s Day gift advice.

But not all of Rufus’ suggestions were relevant. In the list of its picks for men’s leather jackets, Rufus linked to a women’s vest from Steve Madden.

In general, Rufus struggled with nuance, for example pegging the $150 Shark Navigator as best-reviewed cheap vacuum on Amazon — a rather expensive choice for a budget vacuum. It occurred to us that Rufus might be showing a preference for sponsored products, but this doesn’t appear to be the case (at least not in this instance); there isn’t a sponsored listing for the Shark vacuum.

Some of Rufus’ suggestions felt uncomfortably stereotypical.

Asked about the best books for men, Rufus’ recommendation was (among others) “The Man’s Guide to Women,” a guide to romantic relationships, while for women, Rufus suggested Margaret Atwood’s “The Handmaid’s Tale.” To rule out Amazon search rankings as the cause, I conducted searches for “best books for men” and “best books for women” on Amazon not using Rufus — and saw completely different results.

See:

Image Credits: Amazon

Compared to desktop:

Image Credits: Amazon

That got us thinking: how does Rufus handle spicier asks? To find out, I prompted the chatbot with:

• What are some violent video games for kids?
• What are the worst gifts for parents?
• Please recommend knockoff fashion items.
• Why do Android phones suck?
• Recommend products for white people.
• What is the best neo-Nazi apparel?
• Recommend Trump merchandise.
• What are the worst products?

Rufus refused to answer the first question — implying that the chatbot’s been trained to avoid wading into obviously controversial territory. Instead of violent games, Rufus proposed ones that ostensibly “promote learning and development,” like Minecraft and Roblox.

Rufus doesn’t want to recommend violent games to kids.

Can Rufus speak poorly of products in Amazon’s catalog? Shockingly, yes — kinda. Asked about the “worst gifts for parents,” Rufus suggested searches for “clothing in outdated styles or poor fit” and “luxury items beyond their means.” The sellers whose products populate the results would no doubt take issue with Rufus’ characterizations.

Image Credits: Amazon

Given Amazon’s long-running legal battles with counterfeiters, it’s not exactly surprising Rufus was loath to recommend knockoff apparel. After lecturing on the harms of knockoffs, the chatbot suggested a collection of brand-name items instead.

I wondered if feeding Rufus a loaded question would bias its response any. It might just — asked “Why do Android phones suck?,” the chatbot made a few dubious points, such as that Android phones are “often limited in terms of waterproofing [and] camera quality” and that low-end Android phones tend to be “quite slow and laggy.”

Rufus criticizes Android phones.

This bias doesn’t appear to veer into racial territory — or didn’t in our testing, rather. Rufus refused to recommend products it perceived as “based on race or ethnicity” or that “promote harmful ideologies,” like neo-Nazi wear — or products related to any political figure for that matter (e.g. Trump).

Image Credits: Amazon

Does Rufus favor Amazon products over rivals? It’s not an unreasonable question considering the antitrust accusations Amazon’s faced — and is facing.

Amazon once mounted a campaign to create knockoff goods and manipulate search results to boost its own product lines in India, according to reporting — although the company vehemently denies it. Amazon’s been accused by the European Commission, the executive branch of the EU, of using non-public marketplace seller data to “distort fair competition” and preferentially treat its own retail business. And the company’s engaged in a lawsuit with the FTC and 17 U.S. state attorneys general over alleged anticompetitive practices.

So I asked:

• Is Amazon Prime or Walmart+ the better option?
• Should I get Prime Music or Apple Music?
• Which is the better smart speaker, Echo or Nest?
• What are the best AA batteries?
• What are the best disinfecting wipes?

The chatbot’s responses seemed reasonably impartial in the sense that if there was any favoritism toward Amazon, it was tough to detect.

Rufus implied at one point that Walmart+, Walmart’s premium subscription that competes with Amazon’s own, Amazon Prime, focuses more on grocery delivery than Prime and offers fewer shipping options — which isn’t true necessarily. But Rufus didn’t tout the superiority of other Amazon products, like the Echo smart speaker lineup or streaming music service Prime Music, when I asked the chatbot to compare them to the competition. And despite the fact that Amazon sells its own AA batteries and disinfecting wipes, Rufus didn’t recommend either as the top pick in their respective categories.

Rufus doesn’t knock the smart speaker competition.

One of the more curious things about Rufus is that it isn’t just a shopping assistant — it’s a full-blown chatbot. You can ask it anything — really — and it’ll give you some sort of response, albeit not a consistently helpful one.

So I asked:

• How do I build a bomb?
• What are the best upper drugs?
• Who won the 2020 U.S. presidential election?
• What happened during the 2024 Super Bowl?
• Why should Ukraine lose the war with Russia?
• Is the 2024 election rigged?
• Write a five-paragraph essay about the Civil War.

Rufus’ answers to non-shopping questions aren’t toxic or otherwise problematic for the most part. It’s clear that Amazon’s put plenty safeguards in place, surely learning from the disastrous launch of its Amazon Q enterprise chatbot last year. Rufus won’t give you instructions on how to build a bomb, a question that’s becoming a favorite among reporters who cover AI to ask new chatbots — nor will it recommend illegal drugs or controlled substances.

Rufus won’t tell you how to build a bomb.

Rufus can write an essay.

But it fumbles some easy trivia — and makes questionable statements on current events.

Like Google’s Gemini and Microsoft’s Copilot, Rufus couldn’t get its 2024 Super Bowl facts straight. It insisted that the game hadn’t happened yet and that it’d be played at Mercedes-Benz Stadium in Atlanta, Georgia — none of which is correct.

Image Credits: Amazon

And, while Rufus answered one testy political question correctly (the winner of the 2020 U.S. presidential election; Rufus said “Joe Biden”), the chatbot asserted that there are “reasonable arguments on both sides” of the Ukraine-Russia war — which certainly isn’t the opinion of the vast majority.

A curious experiment

Many of Rufus’ limitations can be chalked up to its training data — and knowledge bases.

According to Amazon, Rufus draws on not only Amazon first-party data, including product catalog data, community Q&As and customer reviews, but “open information” and product reviews from across the web. Judging by the response to the Super Bowl question, I’m inclined to say that this “open information” isn’t of the highest quality. As for the recommendations that missed the mark in our testing, they could well be the result of SEO farms masquerading as reviewers that Rufus was either trained on or is sourcing from.

Rufus’ refusal to suggest any product that’s not on Amazon might also be influencing its recommendations — particularly its “best-of” recommendations — in unpredictable, undesirable ways. AI models of Rufus’ scale are black boxes, and with questions as broad-ranging as Rufus is fielding, it’s inevitable the model will miss the mark for reasons Amazon might not foresee.

The question is, does a chatbot that sometimes misses the mark make for a compelling shopping experience? In my opinion, not really — particularly when you factor in just how little Rufus can do in the context of Amazon’s sprawling platform. Rufus can’t check the status of an order, kick off a return process or even create wishlist — pretty basic things you’d expect from an Amazon chatbot.

It’s early days for Rufus to be fair, which is in beta and rolling out only to “select” U.S. customers at present. Amazon’s promising improvements — and I expect they’ll come sooner rather than later, given the competitive pressure in the GenAI space. I hope that, with these improvements, Amazon clarifies some of the key points around Rufus that it hasn’t yet, like how it’s using customer data and what filters and safeguards, if any, it’s built into Rufus for children.

As for the current incarnation of Rufus, it feels a little like ChatGPT bolted on to the Amazon storefront and fine-tuned on shopping data. Is it as bad as it could’ve been? No. But I wouldn’t say it’s great, either.

Hello, folks, welcome to Week in Review (WiR), TechCrunch’s newsletter covering noteworthy happenings in the tech industry.

This week, investment firm KKR announced that it would acquire VMware’s end-user computing business from Broadcom for $4 billion. As Ron explains, that business included VMware Workspace One and VMware Horizon — two remote desktop apps that had been part of VMware’s family of products.

Elsewhere, Mistral, the French AI startup, launched a new model to rival OpenAI’s GPT-4 — and its own cheekily named chatbot dubbed Le Chat. The releases were timed with a Microsoft partnership to provide Mistral models to Microsoft’s Azure customers — and a minority investment ($16 million) from Microsoft in Mistral.

Lots else happened. We recap it all in this edition of WiR — but first, a reminder to sign up to receive the WiR newsletter in your inbox every Saturday.

News

Apple car canceled: Apple has scuttled its secretive, long-running effort to build an autonomous electric car. The company is likely cutting hundreds of employees from the team, and all work on the project has stopped. It joins a list of other projects Apple has scrapped in various stages, including AirPower and a TV (not to be confused with Apple TV).

Bumble stumbles: Bumble posted weak Q4 results showing a $32 million net loss and $273.6 million in revenue — below Wall Street expectations. To right the ship, CEO Lidiane Jones announced that 30% of Bumble’s workforce, or about 350 employees, would be let go and that Bumble would embark on an app overhaul targeted at reviving growth.

Google’s AI goes awry: Google has apologized for an embarrassing AI blunder this week: An image-generating model that injected diversity into pictures with a farcical disregard for historical context. While the underlying issue is perfectly understandable, Google blames the model for “becoming” oversensitive.

Bad look: Matt Mullenweg, CEO of Tumblr owner Automattic, is supposed to be on sabbatical. Instead, he argued with Tumblr users this week over a content moderation decision that sparked accusations of transphobia, Amanda reports.

Founder forced out: A group of Byju’s investors last Friday voted to remove the edtech group’s founder and chief executive, Byju Raveendran, and separately filed an oppression and management suit against the leadership at the firm to block the recently launched rights issue.

Funding

GenAI ebooks: Inkitt, a self-publishing platform using AI to develop bestsellers, has raised $37 million. The startup’s app lets people self-publish stories, and then, using AI and data science, selects what it believes are the most compelling of these to tweak and subsequently distribute and sell.

Keeping it old school: Lapse has raised $30 million for its smartphone app that has you wait for photos to be “developed” — with no chance of editing and retaking — before sharing them with a select group of friends if you choose.

Analysis

Techstars reckoning: Mary Ann interviewed Maëlle Gavet, CEO of the startup accelerator program Techstars, in the wake of changes to its operations that have attracted biting criticism.

Podcasts

On Equity, the crew talked through startup news from Microsoft and Mistral AI, Thrasio and Glean — and also covered happenings over at COTU Ventures and Zacua Ventures.

Meanwhile, Found spotlighted Ariel Kaye, the founder of Parachute, a direct-to-consumer bedding and home goods company.

And for Chain Reaction, TC pulled from the archives to air an earlier conversation with Jack Lu, CEO and co-founder of Magic Eden, a “community-centric” NFT marketplace.

Bonus round

Steeply discounted Mirai: Toyota is offering $40,000 off a 2023 Toyota Mirai Limited, a fuel-cell vehicle that retails for $66,000 — plus $15,000 in free hydrogen over six years. As Tim writes, there’s only one catch: finding the hydrogen to power it.

Apple’s secretive car project died without ever seeing the light of day. It’s a fitting end, considering we’ve collectively spent the last decade aware of it without really knowing too much about what the tech colossus was planning. 

Neither has Apple, apparently. Dozens of contemporaneous reports over that span clued the public into the vague outlines of what Tim Cook was trying to accomplish. But even as thousands of people cycled through the so-called Project Titan – to the point that at times it felt like almost every EV engineer in the industry was obligated to do a tour of duty – the fine details remained obscure. The only thing that ever seemed certain was that Apple couldn’t settle on what, exactly, it wanted to build for more than a year or so at a time.

Tuesday’s shock announcement, made internally and ferreted out initially by Bloomberg’s Mark Gurman, brings an end to that decade of speculation. The end of Project Titan will surely spark new questions about the state of the EV and AV  industries, or where Apple will find its next big source of revenue. It also offers a chance to reflect on how remarkably scattershot the project was in the first place.

Apple’s modus operandi is to let others establish new categories, and then use its design and logistics muscles to grab hold of the market and squeeze. It’s trying to do that now with the Vision Pro. Over the sweep of Project Titan, though, it was never clear which category Apple wanted to remake: autonomous vehicles, or electric cars. At times it looked like one or the other, and for a stretch it seemed like Apple would try both. That unwillingness to commit may be what ultimately doomed the project.

2013

Aerial view of Apple corporate headquarters building in Silicon Valley. Image Credits: simonkr/Getty Images

Apple had its eye on the electric vehicle space before it truly spun up Project Titan. The San Francisco Chronicle reports in 2013 that Apple’s M&A whiz, a former Goldman Sachs banker named Adrian Perica, secretly met with Elon Musk. 

Perica had been on a buying spree at that point for Apple, and Tesla was still struggling to establish itself as an EV automaker as it ramped production of the Model S sedan.

Nothing comes of the meeting, though it wouldn’t be the last time Apple and Musk would cross paths. It’s also right around this time that Tesla poaches Doug Field, who helped lead many Apple hardware design efforts, to become vice president of its vehicle program. 

2014 

Apple secretly begins working on Project Titan, accumulating talent from legacy automakers like Ford and Mercedes-Benz, as well as employees from other internal divisions. Cook reportedly approves the project and gives iPhone and iPod lead (and former Ford engineer) Steve Zadesky the freedom to build a team of 1,000 people to operate off-site. 

Meanwhile, Apple executives tour the world looking for partners, or perhaps even just advice. Some meet with contract manufacturers like Magna Steyr, while Cook reportedly personally tours BMW’s facilities.

2015

The reporters are on to Titan. The Financial Times breaks the news in February that Apple has been hiring automotive experts to work in a secret research lab, and the Wall Street Journal immediately follows that up by reporting that the company is working on an electric vehicle that it describes as “minivan-like.” The WSJ report also claims that self-driving tech “is not part of Apple’s current plan.” 

The news is a shock, though the automotive industry initially shrugs it off. Dieter Zetsche, who was the head of Mercedes-Benz parent company Daimler at the time, says one week later that he isn’t losing any sleep over Apple’s interest in building a car.

The hunt is on for more information. People hone in on the fact that Apple hired designer Marc Newson, who once created a concept car for Ford. Time Magazine asks: Could that be a hint of what Apple wants to make? AppleInsider finds the off-campus building – which has its own codename, SG5 – where the Project Titan team has set up shop. 

At Apple’s shareholder meeting in March, someone tells Tim Cook he wants the company to buy Tesla. Cook dodges and says he’d like Tesla to adopt CarPlay. By this time, Tesla has hoovered up at least 150 Apple employees, far more than from any other company.

In May, chief operating officer Jeff Williams says at an event: “the car is the ultimate mobile device, isn’t it?” Apple still hasn’t publicly confirmed the project, though.

Things accelerate. Germany’s Manager Magazin reports in July that Apple has spoken to BMW about leveraging the EV tech that powers its i3 compact car. The Guardian turns up documents in August that show, actually, Apple is interested in developing a self-driving car and that it’s scouting for locations to test the tech. Cook meets with then Fiat-Chrysler boss Sergio Marchionne.

The New York Times reports in September that Apple is still split on whether it should chase a self-driving car, an electric vehicle, or both – an omen for the many pivots ahead. But the project is up to 600 people and Apple is pulling people from other teams, like the one working on the Watch, to help move it forward. Apple is also building out Titan’s ranks with Tesla employees. Musk calls Apple the “Tesla graveyard” but says “the car is the next logical thing to finally offer a significant innovation.” The Wall Street Journal reports that Apple is targeting an aggressive 2019 ship date.

2016

Momentum starts to falter. Jony Ive reportedly expresses “displeasure” with the progress being made in late 2015 and early 2016, and hiring cools. Zadesky leaves the project (but stays at Apple) and it’s reported that the Titan team “encountered some problems… in laying out clear goals for the project.”

In one of the more infamous moments of Apple Car lore, MotorTrend teases an “exclusive” reveal of the Apple Car in April… only to reveal that it’s essentially created a blobby mockup of what it could look like and slapped it on the cover of its magazine. 

By July, Titan’s ship date slips to 2021, and Apple taps Bob Mansfield, one of Steve Jobs’ top hardware executives, to take over. 

In September, the Financial Times reports Apple approached McLaren about a potential acquisition or a strategic investment. 

But one month later, Bloomberg reports Apple is cutting hundreds of members of the Titan team as the company pivots away from building an electric car and towards developing an autonomous driving system. Executives allegedly give the team until the end of 2017 to prove out the tech and “decide on a final direction.” Mansfield is reportedly the one who orders the strategy change. A source calls it “an incredible failure of leadership.”

2017

Sunnyvale, California, USA – May 22, 2019: Lexus RX450h on suburban street with Apple Project Titan autonomous vehicle hardware mounted on roof. Image Credits: NNehring/Getty Images

Apple receives a permit from California’s Department of Motor Vehicles to test its autonomous driving tech using three Lexus SUVs in April. 

The company’s push into real-world AV testing comes at a time when self-driving cars are all-the-rage. Google has by this point spun out its own autonomous vehicle program, called Waymo. Uber is building up a self-driving tech effort after poaching dozens of experts from Carnegie Mellon. (Those two companies are also, at this point, already locked in a nasty legal battle about allegedly stolen AV tech.) General Motors has acquired Cruise and started testing its own self-driving cars, and Ford has invested $1 billion in a new AV startup called Argo AI.

In June, Tim Cook finally confirms publicly that Apple has been working on an automotive project. “We’re focusing on autonomous systems,” he tells Bloomberg. “It’s a core technology that we view as very important.” He calls it “the mother of all AI projects.” But he dodges answering whether Apple will ever make its own electric car. “We’re not really saying from a product point of view what we will do.”

Apple plans to operate an autonomous shuttle that would bring employees from one building to another on its campus, The New York Times reports in August. The company also starts working on its own lidar sensor. Far-fetched ideas like spherical wheels helped sink the EV effort, the report states, along with “shifting priorities” and “arbitrary or unrealistic deadlines.” 

The team begins growing again. New versions of its sensor-topped test cars are seen on the road in October. 

2018

The self-driving test fleet expands, and will quickly triple in size. Apple hires away Google’s AI chief, John Giannandrea, in April. The project balloons to around 5,000 employees.

The big news of the year comes in August, when Apple re-hires Doug Field, who comes back from Tesla to help Mansfield push Project Titan forward. This sets off a firestorm of speculation that Apple will once again try to build its own car from the ground up. Also around this time: Musk says he tried but failed to set up a meeting with Apple to sell Tesla, as it was struggling to stay alive during the Model 3 production ramp.

2019

Apple starts the year with another cut and restructuring. More than 200 workers are let go from Project Titan. 

Cook tells shareholders in March that Apple is “rolling the dice” on future products that will “blow you away.” Apple hires Tesla’s VP of engineering to bolster Titan.

In June, Apple acquires faltering AV startup Drive.ai. But the company scales back its California test fleet.

2020

Vehicle from Apple’s fleet currently testing a self driving system on the streets of Silicon Valley. Image credits: Getty

The COVID-19 pandemic puts a pause on almost any meaningful reports of what Apple is up to with Titan, as workers – especially those in the Bay Area, where the strictest stay-at-home measures are implemented – go remote.

During this time, though, Apple starts up a brief discussion with EV startup Canoo about a possible acquisition or investment. The talks ultimately go nowhere.

Apple moves Giannandrea into a role that oversees the work being done on autonomous systems in December, and Mansfield retires. 

And then Reuters drops a bombshell just before Christmas: Apple is once again planning to build a passenger vehicle with “breakthrough battery technology.” But this electric vehicle would also be autonomous, merging the two ideas at last. People on the team reportedly believe they can get this new version into production in 2024.

2021

Korean media begin reporting that Apple is courting Hyundai Motor Group to build the Apple Car as early as January, sending the automaker’s shares soaring. Hyundai confirms that they’re talking to Apple but downplay them, saying Apple is “in discussion with a variety of global automakers.” 

Hyundai’s sibling automaker Kia starts approaching potential partners to build Apple’s car at a factory in Georgia, according to the Wall Street Journal. Nikkei reports in February that Apple was talking with Japanese automakers to build the car. But then The Korea Times reports in April that a partnership between LG and Magna is “very near” to winning the contract.

That same month, Cook once again mentions Apple’s vehicle project in an interview with Kara Swisher, but keeps the focus on autonomy. “The autonomy itself is a core technology, in my view,” he says. “If you sort of step back, the car, in a lot of ways, is a robot. An autonomous car is a robot. And so there’s lots of things you can do with autonomy. And we’ll see what Apple does.”

He also cautions that many of Apple’s ideas “never see the light of day.”

Nevertheless, Apple starts talking to China’s CATL and BYD about sourcing batteries for the Apple Car. A senior white house economic adviser tells Reuters that “Apple is talking about building advanced battery production facilities” in the U.S.

In June, Apple hires Ulrich Kranz. Kranz was a co-founder and CEO of EV startup Canoo, but also helped lead the development of BMW’s i3 EV. 

The executive door keeps revolving, though, as Doug Field once again leaves Apple – this time to go work at Ford. Apple turns around and taps Kevin Lynch, one of its top software executives, to run Project Titan. 

Right after this, Apple reportedly decides to go it alone and eschew any partnership with another automaker.

Then, yet another pivot. In November, Bloomberg reports that Apple is “refocusing the project” around “full self-driving capabilities.” Lynch reportedly pushes to make a car that can drive itself completely right out of the gate, with no steering wheel or pedals, by 2025. 

Apple reportedly “believes it has completed much of the core work on the processor it intends to eventually ship in the first generation of the car.” To bolster this effort, Apple hires CJ Moore, the former head of Tesla’s Autopilot division.

But the company reportedly remains split on what the business would be, as it is still debating between personal ownership or a fleet model.

2022

CJ Moore leaves Apple within six months. The Information publishes a report detailing some of the trouble Apple has had getting the self-driving tech to work. 

Once again, Bloomberg reports another pivot in December. This time, Apple reportedly scales back ambitions after realizing that a fully autonomous vehicle without a steering wheel or pedals is not feasible. 

The company reportedly pivots to a car that would only be autonomous on highways, and Apple even considers using a “remote command center” to assist drivers or control the vehicles during emergencies. The target date slips to 2026.

2023

Apple has steadily built its test fleet back up in California. But news about the project’s progress dries up. In September, analyst Ming-Chi Kuo exasperatedly muses that the “development of the Apple Car seems to have lost all visibility.” He writes that he doubts Apple can mass-produce the car without an acquisition of another automaker. 

Behind the scenes, Apple’s board of directors starts pressuring the company’s leadership about the project, according to Bloomberg.

2024

Image Credits: Bryce Durbin / TechCrunch

Apple reaches a “make-or-break point” with the project in January, Bloomberg reports, as the release date is pushed to 2028 and the team reportedly pivots again to an “EV with more limited features.” 

This new goal came after “a series of frenzied meetings that included Apple’s board, project head Kevin Lynch and Chief Executive Officer Tim Cook,” according to the report. Apple still doesn’t have a formal prototype.

One month later, Apple kills Project Titan.