.png)
3 hours ago
2
Hello, and welcome back to TechScape. This week in tech: Elon Musk and Donald Trump flood the zone and deploy brinkmanship as a negotiating tactic; US Immigration and Customs Enforcement learns search engine optimization amid arrests and deportations; and Spotify tries to soften its algorithmic image with human-centric public relations. Thank you for reading.
Flood the zone: Trump has executive orders; Musk has Doge
Donald Trump has issued a record number of executive orders since his presidency began: ending birthright citizenship, banning gender transitions for anyone under 19, pardoning the rioters of the January 6 attack, and more. Elon Musk, the world’s richest man in charge of the “department of government efficiency”, has raided an equally dizzying swath of federal agencies with the stated goal of “slashing waste, fraud, and abuse”. Among the half-dozen bureaus are the Consumer Financial Protection Bureau, Department of Education, Department of Labor and, most viciously, the US Agency for International Development (USAid). Trump and Musk are doing their utmost to “flood the zone” – a tactic that the former Trump administration strategist Steve Bannon has touted as one that will purposefully overwhelm the opposition and the media. Bannon is right; it’s tough to keep up.
Trump and Musk share a flair for brinkmanship. They force their opponents to accept a version of what they want by forcing negotiations to the point of mutually assured destruction, then retreating. Trump declared he would use executive orders to impose 25% tariffs on all goods imported to the US from Mexico and Canada, enormous penalties to exact from the country’s two biggest trading partners that threaten to send all three economies into tailspins. However, Trump delayed the tariffs a month after Canada and Mexico each agreed to staff up their borders with the US with 10,000 troops.
Musk preaches that he’s dissolving USAid. He might yet – the agency’s name has been taken off its door – but the likelier outcome seems like a diminished version of the agency will be folded into the state department. Secretary of state Marco Rubio has declared himself head of the agency as workers are locked out. Musk’s fingerprints were visible on the “fork in the road” email received by all 2 million employees of the US government. Will they all resign, evacuating and disabling the government? Extremely improbable. Does a partial buyout of a workforce achieve the same result as a layoff? Yes.
Musk has used this playbook many times before, most notably with X, née Twitter. He yanked office leases so fast that the company failed to pay rent it still owed. He fired about three-quarters of the workforce and left the question of severance to the courts after some ex-employees sued. It seems inevitable Musk will owe less after negotiating payouts via lawsuits. The edgy approach hasn’t always worked out for Musk, though. With X, he tweeted often about how bad the social network was to gain attention, threatened to buy it, signed a contract for the $44bn purchase, then did his best to pull out of the agreement. He was forced to go through with it, though, and acknowledged X was overvalued in a recent interview. When a judge invalidated his $56bn pay package at Tesla, he said he would consider walking away from the company without the compensation. The judge didn’t blink and ruled his pay excessive a second time. Musk has neither left Tesla, nor does he have an existing pay package.
Silicon Valley software companies like to say they move fast and break things, iterating their products in service of innovation. I would call what Elon Musk does moving so fast he might break everything. It seems to be one of the things Trump likes best about him. The president himself is not so different. Remember when the US almost went to war with Iran? He called off airstrikes while the planes were in the air.
Musk’s next target seems to be Medicare and Medicaid. What will break there, and who will it break?
Read about who is helping Elon Musk lay waste to the US government here.
The logical end of SEO: a mirage of deportations in Google search results
There are several ways to game Google search to boost a website to the top of the results page, the most valuable real estate on the internet. In fact, a whole field is built around it called search engine optimization, or SEO. Google’s algorithm works by looking at various factors on a webpage to determine if it is relevant and authoritative. Government web domains already get authoritative bonus points.
SEO tactics are in the hands of every middling recipe blogger in the United States, why wouldn’t the government know about them? US Immigrations and Customs Enforcement (Ice) updated thousands of press releases on 24 January, making old press releases seem new in Google search results, befuddling a longtime immigration lawyer’s search for enforcement actions.
My colleague Dara Kerr reports:
News of mass immigration arrests has swept across the US over the past couple of weeks. Reports from Massachusetts to Idaho have described agents from Ice spreading through communities and rounding people up. Quick Google searches for Ice operations, raids and arrests return a deluge of government press releases. Headlines include ICE arrests 85 during 4-day Colorado operation, New Orleans focuses targeted operations on 123 criminal noncitizens and, in Wisconsin, ICE arrests 83 criminal aliens.
But a closer look at these Ice reports tells a different story.
That four-day operation in Colorado? It happened in November 2010. The 123 people targeted in New Orleans? That was February of last year. Wisconsin? September 2018. There are thousands of examples of this throughout all 50 states – Ice press releases that have reached the first page of Google search results, making it seem like enforcement actions just happened, when in actuality they occurred months or years ago. Some, such as the arrest of “44 absconders” in Nebraska, go back as far as 2008.
All the archived Ice press releases soaring to the top of Google search results were marked with the same timestamp and read: “Updated: 01/24/2025”.
Read the full story here.
Spotify plays up the human side of its recommendations while going big on AI features
Björk recently called Spotify “probably the worst thing that has happened to musicians. The streaming culture has changed an entire society and an entire generation of artists.”
How to counter such bad publicity as the doyen of experimental music deriding your company? Spotify was ready, though. The company had been on the charm offensive in advance of its sunny earnings report last week, when it posted its first-ever full-year profit. CEO Daniel Ek said he saw “potential for transformative shifts in music discovery” in recent developments in AI.
In mid-January, the Wall Street Journal published a profile of Spotify’s eccentric head of editorial, with the headline and subhead: The Playlist Power Broker Who Makes or Breaks New Artists: Spotify is known for its algorithmic recommendations, but Sulinna Ong brings a human touch to finding new hits. Ong told the paper she spends three hours a day lying on the floor listening to new songs. How whimsical, how human! She positioned herself as the anti-android soul within the mechanical beast: “AI machine learning is amazing at parsing large data sets, but when there is no data, for example, on a new release, on a new artist, what does it do?”
I doubt how much humanity Ong is able to inject. Spotify’s playlists are certainly big tastemakers, but serving 640 million users doesn’t happen with loving curation by hand, and the app is placing less emphasis than ever on the human touch. It algorithmically generates playlists based on your music taste and the activity you search for, eg “Running Mix”, “Daily Drive”, or “Feel Good Cleaning Mix”, humans not included. The streamer laid off about 1,700 employees in December 2023, including many of its human playlist makers. Four months later, it rolled out the AI playlist generator, which was probably trained on the labor of those curators, that allows users to populate playlists based on text prompts like a musical ChatGPT.
Spotify has positioned Ong in the pages of WSJ like a bishop on a chess board. Every interview with a current employee of a tech company that makes it into print happens with a predetermined goal in mind. An executive as senior as Ong won’t go near a journalist without the blessing of a dozen public relations professionals and Spotify executives. Such sit-downs are selectively and strategically deployed in service of the narrative those executives believe will drum up consumer support and dollars. She’s not on the payroll to sit for photoshoots, though perhaps she can listen to an hour of new songs while posing.
What tune is Spotify singing, then? The company is emphasizing the humanity of its recommendations even as its financial and product decisions scream AI. The public doesn’t like AI going near art; hundreds of artists in music and other disciplines have spoken out against generative AI and accused its developers of stealing their work. As a tech company, though, Spotify wants AI to refine its recommendations at the scale of hundreds of millions. That’s where its money is going; that’s what the engineers are working. As a music company, it needs the semblance of a human touch. Enter Sulinna Ong, its persuasively human mascot.
I expect more companies will follow the same public relations playbook, spending big on AI but emphasizing the small tweaks humans make. Even Björk herself is not immune. She trashed Spotify in an interview that was mostly about promoting her new concert film on Apple TV+.
