Back in the early 2000s when I attended a lot of tech conferences, there was a common and usually regretful discussion: is tech now a mature industry? Even though Facebook, YouTube, the iPhone, Uber, and Bitcoin were all in the future, there was already concern that the “old” insurgent Silicon Valley was somehow lost, and that the startup ethos — often romanticized — was giving way to regularized processes, HR departments, and boring stuff like earnings calls and firm shipping deadlines. If people were sad in 2002, what must they make of today’s environment? We see big funding for food delivery and other gig work companies, the many consequences of ad-funded software, and no U.S. tech companies of wide impact started in more than a decade (Uber launched in 2009).
1) No tech company with work to do wants to buy back stocks with its spare cash: engineering the share price rather than investing in growth worries me and a lot of other people a great deal. I can’t remember which business professor said it (probably several), but the paraphrase was that “buying back your stock tells me you’ve run out of fresh ideas.” At the same time, when you have the cash pile Apple does, there’s money for lots of things, including acquisitions, buybacks, and new product launches.
2) Any firm that tries to protect its business model primarily by lobbying sends a bad signal. We like the status quo, it seems to say, and we want to protect it with favorable legislation rather than having to compete and win in the market.
3) Buying big corporate headquarters sends more negative signals. Salesforce and Apple both make statements here. Amazon’s bake-off approach to “HQ2” created negative optics. To their credit, Microsoft is downsizing its physical office space and eschewing the “edifice complex.” Netflix spent big on movie production facilities, but that’s essentially a factory in their industry.
By far the biggest marker of post-mature status is geopolitical activity. The Trump administration took action in this domain (aimed at TikTok and Huawei) while the Biden administration is making noises about Amazon and Facebook. Beyond US presidential initiatives, several tech companies have had major geopolitical episodes in the past five years:
-Facebook proved to be a decisive “threat vector” for foreign influence in the 2016 elections
-Twitter became a key tool of the Trump presidency
-Google left China
-YouTube is used by extremist groups that post misinformation and violent content for recruiting purposes.
The most recent and farthest reaching actor in the geopolitical entanglement with the tech sector is the Chinese government. Alibaba’s Jack Ma (worth a reported $50 billion) actually disappeared for three months last November on the eve of his Ant Group's IPO; a company executive said he is now focusing on hobbies. Didi (a Chinese ride-hailing app company) went public in the US only to have its home-market operating status reined in soon thereafter, dropping its equity value substantially. TikTok parent ByteDance did not go public as many expected; instead, the Chinese government took a board seat on a ByteDance operating company, complicating the US app’s promises that stateside user data is not shared with the core team in China.
But blocking big gains in personal wealth, while ideologically coherent, is not the only thrust of current efforts. Last week the CAC (Cyberspace Administration of China) posted a draft of regulations designed to apply to “Internet Information Service Algorithms” and there is a _lot_ in the document. My first thought upon reading it: how will western governments respond to a regime widely characterized as authoritarian ostensibly taking the lead on protecting citizens from algorithmic abuses?
Here’s the document: http://www.cac.gov.cn/2021-08/27/c_1631652502874117.htm#. I’m relying on Google Translate and will gladly correct any readings that I get wrong given my reliance on an imperfect tool. All bold-face emphasis is mine.
The document gets very granular very quickly: article 2 states that “The application algorithm recommendation technology mentioned in the preceding paragraph refers to the application of algorithmic technologies such as generating synthesis class, personalized push class, sorting selection class, retrieval filter class, scheduling decision class and so on to provide users with information content.” This looks like it could be extended to include not only adtech but social media, ride hailing, meal delivery, discussion threads, and tutoring (another recent point of emphasis).
Article 6 paints with a very broad brush: I’m sure phrases like “mainstream value orientation,” “positive energy,” and “economic and social order” are carefully chosen and carry great weight. The paragraphs also appear to imply censorship of a wide range of content.
“Algorithm recommendation service providers shall adhere to the mainstream value orientation, optimize the algorithm recommendation service mechanism, actively disseminate positive energy, and promote the upward improvement of algorithm application.
Algorithm recommendation service providers shall not use algorithm recommendation service to engage in activities prohibited by laws and administrative regulations, such as endangering national security, disrupting economic and social order, and infringing on the legitimate rights and interests of others, and shall not use algorithm recommendation services to disseminate information prohibited by laws and administrative regulations.”
Article 8 seemingly prohibits the dopamine-pump model that currently powers Facebook and TikTok: “Algorithm recommendation service providers shall regularly review, evaluate and verify the algorithm mechanism mechanism, model, data and application results, etc., and shall not set up algorithm models that induce users to indulge or consume in high amounts that violate public order and good customs.”
Article 14 appears to call for algorithmic transparency. “The provider of algorithm recommendation service shall inform the user of the situation of providing algorithm recommendation service in a significant way, and publicize the basic principle, purpose intention and operation mechanism of algorithm recommendation service in an appropriate way.”
Article 15 seems to endorse an opt-in rather than the dominant and delimited opt-out model.
“Algorithm recommendation service providers shall provide users with options that are not specific to their personal characteristics, or provide users with convenient options to close algorithm recommendation services. If the user chooses to close the algorithm recommendation service, the algorithm recommendation service provider shall immediately stop providing the relevant service.
The algorithm recommendation service provider shall provide users with the function of selecting, modifying or deleting user tags used for algorithm recommendation services.
Users who believe that the application of the algorithm of the algorithm recommendation service provider has a significant impact on their rights and interests have the right to require the algorithm recommendation service provider to explain and take corresponding improvements or remedial measures.”
Might that include an EU-like right to be forgotten?
Article 17 sounds like it is directed at rideshare and meal delivery.
“Where algorithm recommendation service providers provide work scheduling services to workers, they shall establish and improve platform order allocation, remuneration composition and payment, working hours, rewards and punishments and other related algorithms, and fulfill the obligations of protecting workers' rights and interests.”
Article 18 would seem to limit differential pricing of consumer goods, especially when it is based on user behavior (US airlines, take note):
“If an algorithm recommends that a service provider sells commodities or provides services to consumers, it shall protect the legitimate rights and interests of consumers, and shall not use algorithms to carry out unreasonable differential treatment in trading conditions such as transaction prices and other illegal acts according to the preferences, trading habits and other characteristics of consumers.”
Does article 20 mandate state oversight of the actual code base?
“Algorithm recommendation service providers with public opinion attributes or social mobilization ability shall fill in the name, service form, application field, algorithm type, algorithm self-evaluation report, content to be announced, etc. through the Internet information service algorithm record system within 10 working days from the date of service delivery.”
For all the breadth of these proposed measures, the financial consequences are light: fines range from 5000 yuan to 30,000 yuan ($773 to $4639). At the same time, even the EU has not proposed, much less implemented, any sort of public algorithmic oversight, an opt-in model for collection of user data, or prohibitions on hooking kids on apps while they are young: providers “shall not use algorithmic recommendation services to induce minors to indulge in the Internet.” Just this week, according to Reuters, "China has forbidden under-18s from playing [online] video games for more than three hours a week, a stringent social intervention that it said was needed to pull the plug on a growing addiction to what it once described as 'spiritual opium'."
I’m not a professional China-watcher by any stretch but I do know that the government strictly surveils the citizenry, in part via social media trace data. How do consumer protections relate to the ability of government to collect user data from Internet companies? How do these regulations relate to the Chinese government’s efforts to build its non-consumer tech sector: chip-making, autonomous vehicles, advanced manufacturing, and others?
In answer to the question posed by the newsletter’s title — “what comes after maturity?” — the answer most certainly is “regulation.” The next few years will be most interesting as the tech sector on most every continent will face state scrutiny, litigation, legislation, and taxation. That idealized garage startup is a distant memory.