Ai

I saw that Fanqie Novel was embroiled in another controversy about AI-written content in the past couple of days. My initial reaction wasn’t surprise, but rather a sense that this issue will eventually come to the surface. Considering the combination of free platforms, the pressure for daily updates, and algorithmic distribution, it is almost inevitable that authors will turn to AI to supplement their content capacity.

But honestly, I have a consistent feeling about many books on Fanqie: they are readable, and even the first hundred chapters are often quite good. However, the further along you get, they tend to be nothing more than tropes and speed, lacking that inherent power that classic Xuanhuan or cultivation novels possess. This ‘power’ is hard to explain—it’s probably that feeling when you know it’s a bit over-the-top/cringey (“zhong er”), but you still want to follow the characters all the way through.

It’s something worth noting, and this isn’t meant to criticize the platform. Fanqie being completely free definitely attracts a lot of readers; there’s no arguing that point. However, for someone like me whose taste was spoiled early on by Tian Can Tu Dou, Wo Chi Xi Hong Shi, Er Gen, and Chen Dong, AI can match the output volume, but it cannot replicate the flavor/quality.

In the last few months, while writing code using tools like Claude or Codex, my most striking realization wasn’t that “programmers are obsolete,” but rather that many tasks that used to be given to newcomers for practice can now generate a basic first draft themselves. Whether it’s creating a scaffold, adding several tests, or making small modifications on the fly—after running through these operations, the speed is genuinely fast, so fast it feels almost bittersweet.

For someone like me, who graduated ten years ago, frankly, this is more about increasing efficiency. Because I generally know where it’s reliable and where it isn’t; where something looks functional but actually has pitfalls hidden further down the line. But for fresh graduates, this topic isn’t so straightforward. AI isn’t just here to take over a few hours of manual labor; it feels more like it is compressing the traditional path of how a newcomer goes from zero knowledge to proficiency. This is also why I wanted to write about it separately.

Stunned. / Dumbfounded.

The official ChatGPT side doesn’t make it easy to track tokens and costs directly, so I found a third-party platform and ran a round of similar tasks using GPT-5.4 and GPT-5.5 in Codex, setting the thinking mode to high. The result was very straightforward: simple questions were relatively mild (in terms of cost), with GPT-5.5 being about 30% more expensive than GPT-5.4; however, once complex tasks were involved, the costs shot up to 2.6 times, and both the request count and token consumption increased simultaneously.

My current assessment is very straightforward: this isn’t something that can be decided just because of the statement “5.5 has a higher unit price.” In simple tasks, the cost mainly comes from the unit price; but in complex tasks, what is actually expensive is the entire calling chain (or execution flow). However, looking at it another way, 5.5 does genuinely feel like it’s absorbing your rework costs for you. The model is more willing to think through multiple steps, perform more actions, and check things more thoroughly. Ultimately, the billing isn’t based on a single answer; it’s based on the complete set of actions, which also minimizes the number of back-and-forth cycles required from the human user.

Initially, I didn’t actually plan on testing it. When I came across the news that OpenAI was releasing ChatGPT Images 2.0 on April 21, 2026, my first reaction was just “another image version update.” However, when I checked the Artificial Analysis leaderboard and saw that GPT Image 2 (high) ranked first for text-to-image generation with an Elo of 1332, I felt a bit compelled to test it anyway.

The results are quite impressive; the Chinese output is excellent, it can handle comics, and character/narrative consistency across multiple continuous images has also improved. However, as I tested it further, I felt that what is truly worth discussing this time isn’t “it draws better,” but rather “it starts making things that were previously taken as default truths seem unreliable.” This subject matter is more complicated than a simple leaderboard ranking.

Scrolling through the model updates tonight was genuinely mind-boggling.

My current judgment is straightforward: this round is no longer merely a wave of model releases. It involves three fronts working simultaneously—model capabilities, API pricing, and chip stack ownership. Anyone who focuses on only one of these aspects will likely have a biased view. And it is precisely because these three dimensions are intertwining that the large model sector appears so intensely competitive.

These past couple of days, I came across Anthropic’s Project Glasswing, which is scheduled for release on April 7, 2026. My first reaction was a bit stunned. It wasn’t because another model scored higher, but because it locked the top-tier capabilities into a small circle, initially reserved for defensive players like AWS, Apple, Google, Microsoft, and Linux Foundation.

My own judgment is very direct: This matter is more important than another benchmark record. What frontier AI companies are selling now is no longer just the model itself, but an entire set of access controls—“who gets the capability first, how much capability they get, and what kind of auditing and constraints they have to endure after receiving it.” Models are becoming increasingly like dangerous tools, and the release rhythm is becoming more like issuing licenses.

Seeing Block cut its workforce by 4,000 people out of a group of over 10,000 at the end of February 2026 really shook me. I’ve always worked in financial IT—things like trading pipelines, Hong Kong/US stocks, and system fundamentals. Usually, I’m accustomed to buzzwords like “efficiency improvement,” “automation,” and “cost reduction while increasing efficiency.” But when a fintech company, one so close to money, compliance, and risk control, publicly cites AI as the reason for layoffs, it still hits you hard emotionally.

My current assessment is very direct: the scariest part of AI layoffs isn’t a layoff list in the news one day, but when companies start assuming that “smaller teams can do more work.” This means no backfilling for departures, fewer entry-level positions, and much tighter headcount management. From April 2025 to April 2026, this wind is still blowing in the US; while China hasn’t seen a high-profile, public wave of mass layoffs yet, the quiet squeeze has already begun.

After finishing the article “[Mistaking Hermes for an OpenClaw Alternative, Maybe I Was Biased From the Start]"(/en/p/hermes-openclaw-not-the-same-game/), I went through a round of documentation on both sides. The more I read, the more I felt that to truly see the difference between these two things, just looking at the features isn’t enough; looking at how tokens are consumed is actually more direct.

I’ll state my judgment first.

OpenClaw is by default more like a long-term online workbench; many identities, rules, workspace files, and message constraints naturally persist across conversation rounds, so the base model is usually heavier. Hermes, on the other hand, is noticeably more restrained; much of the context is discovered and injected on demand, and the system prompt deliberately maintains a stable prefix, making it easier to control token usage by default.

Of course, this doesn’t mean that Hermes is necessarily more cost-effective. If you enable the memory provider, skills, sub-agent, and long tool output all at once, it can burn through tokens just as fast. But frankly speaking, these two architectures have not been consuming tokens in the same way since day one.