Today I Learned — 2026-02-10

TIL Series

Today, I wished to write more, but exhausted my capacity (and time) much sooner...

Table of Contents

1. Islamic Philosophy of History

How did medieval Muslim historians write history? This is the question that sets Mustafa Roker to write the following article: Reading Silence: Omission, ʻIbra, and Islamic Philosophy of History.

Referencing Chase F. Robinson's book, Islamic Historiography (Goodreads), Mustafa writes that a medieval Muslim historian deliberately chose what to include/exclude from history. This deliberation was guided by several factors, but the most important one was pedagogical instruction for moral action. The Muslim historian is a key actor, a conscious interpreter who omits, selects, and narrates with the purpose of influencing the reader's judgement.

Mustafa, then, cites two Muslim medieval historians as examples: Ibn al-Jawzī and Ibn al-Athīr. Ibn al-Jawzī, for example, was deliberate about omitting material that lacked authentication or remained too doubtful to warrant transmission. In fact, so conscious was he about his role as a historian that he would exclude certain tales of vice (like a King drinking alcohol or committing adultery), because he was concerned that such tales may be emulated upon and spread depravity in the larger society. Ibn al-Athīr, on the other hand, considered it a waste of space to record minutiae (appointment of a person, slight increase in market price) as he considered them trivial; i.e., practical concerns weighed on his mind.

Mustafa contrasts these Muslim medieval historians with modern historians, such as G.R. Elton (a 20th century empiricist) and Michel-Rolph Trouillot. Elton believed that "facts about the past are simple, discrete, knowable entities which need only be collected in order that a structure called history may emerge." This objectivist portrayal of history as absolute and retrievable by careful ordering of sources was doubted later towards the end of 20th century with the turn towards post-modernism (as an example, perhaps A Thousand Years of Nonlinear History?). History is not merely archival accumulation and narrative reconstruction, says Mustafa: it is a means of cultivating moral discernment. Thus, omission is not a failure of method, but a defining instrument.

Trouillot is cited to show that such omission operates even with secular frameworks. In his account, silences and omissions are shaped by power. Power enters the historical record not only at the level of explicit interpretation, but at the very early stage of what is a 'historical fact', what is preserved, what is later retrieved, and what eventually receives significance.
(I need to read his book: Silencing the Past: Power and the Production of History)

Thus, if a historian cannot separate himself/herself from their work, then one should question the norms authorizing their silence: is it self-constraint, power, bias, or a governing worldview? The article is well-written, so I encourage to read it fully.

Some more thoughts:

2. AI Circular Deals

The famous AI Circular Deals between tech companies is visualized here beautifully: Bloomberg. Unfortunately, it is locked behind a paywall. I found an archive version here: archive.ph; sadly, it does not have the live visualization. But, one can still read through the content.

I won't write the details here; the article is simple enough. An important thing to note here is that circular deals create dependency between these companies. "Circularity can be a winner for all involved if things go well: Company A buys a stake in Company B, giving Company B more money to invest and expand so that, in the end, it needs more of Company A’s products and services." This circularity depends heavily on demand rising. But even if demand rises, competition undercuts pricing. So, how will capital spent on data centers be justified? I promised to write about this in detail later. So, this will have to wait.

3. Platform-to-Cloud: The Economy Transition

This is possibly the best article I read that explains the current cloud economy so coherently: Cloud Capitalism and the AI Transition.

Firstly, the authors give a short overview of the transition from Traditional Fordist firms to Platform Capitalist firms. Initially, firms were both labor and capital intensive. Workers were anchored in geographically anchored production sites. The intensity of work and concentration of workers gave them some power (most were unionized). The businesses were compelled to share rewards of productivity gains with their employees. Later, during the 1970s, these firms faced rising inflation, economic stagnation, and increased competition. A new corporate governance paradigm took hold: the shareholder value. This made firms reorganize around "core competencies" to maximize stock prices and prioritize shareholders' interest over workers. Thus, began the outsourcing of work and retaining only the highest value-added activities (such as knowledge production).

The technological innovations of the IT revolution aided this: it was now feasible to outsource work to the peripheral countries at low cost. The rise of platform capitalism firms began. The strategy of these firms was to capture the market, even if it meant losses in the short-term. Once it solidifies its market dominance, it can live off the "platform rents" not through productive activity, but through monetizing interactions on the platform. Data was crucial for this: the more data captured, the better algorithms can be refined, advertisers can be attracted, and user engagement enhanced. (Platform rents remind me of the inundation of subscription-based models).

The authors state there is a newer transition to a new business model: cloud business model. The Infrastructure-as-a-Service (IaaS) was pioneered by Amazon and rented to external customers through AWS beginning in 2006. Cost savings and flexibility were key reasons to use AWS; client firms could use them on a "pay-as-you-go" model that eliminates the need for heavy upfront investment in "on-premise" systems. Microsoft, Google, and Oracle joined this trend soon-after.

IaaS paved the way for PaaS (Platform-as-a-Service). This was more like managed services (scaling compute and storage based on traffic, new database products that automatically duplicate across geographies, etc). "Narrow" AI that was more task-specific was offered as specialized services. The release of LLMs shifted the focus to scale: compute became more important than relying solely on new innovations in model architecture.

This focus on scale and compute materialized in investments in new data centers. Thus, these cloud firms became asset-heavy — infrastructure reliant with access to energy and water as critical. This also meant that the cloud firms took both software and hardware seriously: a shift from decoupling to recoupling them because it gave a competitive advantage to optimize the hardware for their specialized sophisticated software loads.

Some final points mentioned in the paper that it made more sense for firms to collaborate, than dominate each other. Instead of competing for a larger slice of the pie, they are expanding the "total pie" of cloud capture. And finally, cloud-firms are predominantly enterprise-facing than consumer-facing. This is also a shift from cultivate superficial connection to a wide consumer to a more deep and enduring relationships with enterprises. This is resulting in more and more enterprises (such as Walmart) looking to upgrade to the cloud. Also, because of being geographically constraint (the key asset — data centers — are immobile), it makes sense for the cloud firms to form close connections with federal, state, and local governments. This has given rise to techno-nationalism. Finally, the authors conclude by saying that power has become more concentrated, and it will be difficult for new startups to challenge this.

I rushed through this section, so missed out on painting a more comprehensive picture of the article. I urge you to read it to truly understand the intricacies of the cloud economy. Also, I hope to analyze the business of data center finance in the near future.