@tech1018 Channel Audit: 7,740 Subs, 1,400 Videos, Japanese AI Niche

For Japanese-language tech and AI channels, 7,740 subscribers is solid mid-tier — not a niche micro-channel, but well below the 30K-50K threshold where YouTube starts treating you as a recognizable voice in the algorithm's eyes. There are dozens of Japanese AI commentary channels sitting in the 5K-15K band right now, mostly run by engineers or consultants writing on the side. That's the cluster @tech1018 lives in.

The catalog size is the part that jumps out. 1,400 uploads is an enormous body of work — most channels at this sub count have one to three hundred videos, not four figures. If the count is accurate (and not an artifact of older uploads being counted differently), it suggests this channel has been running for years, possibly with shorter or daily-format content earlier on, before settling into the current Wednesday/Saturday 18:00 schedule the description advertises. Honestly the math is a bit confusing — the scraped total-view figure reads 4,016, which can't be right against a 7,740-sub base and a 1,400-video library. I'd treat that number as broken data and not draw conclusions from it.

The fixed-cadence commitment — 毎週水曜日・土曜日18時公開 — is the kind of discipline marker I'd want more creators to copy. Two scheduled drops a week, named in the description, signals to returning viewers when to come back. The downside, and you see this with a lot of disciplined uploaders, is that pure consistency doesn't fix discovery. You can post a hundred perfectly-timed videos in a row and still not break out if the titles and thumbnails aren't doing work in browse.

The channel's positioning is interesting because it's only partly a media play. Tanaka runs Elvez, an AI development firm in Shibuya, and holds a patent on anomaly-detection-based watchover systems. The YouTube channel is linked into a broader presence — X (@TechSenichiya), Facebook, a note.com blog, and a separate executive blog at hdkworks.com. From outside data alone, this looks like an authority/distribution play feeding back into Elvez's consulting pipeline, not a pure ad-revenue channel. That changes how you'd diagnose growth. A pure creator wants views; an authority play wants the right viewers — and a CTO watching one of his videos is worth more than a thousand passive scrollers.

What I can't see from outside: retention curves, the impressions-CTR loop, exact view distribution across the catalog. The recent uploads I'm looking at scraped with empty titles and zero view counts, which is almost certainly a scraping artifact rather than the channel actually shipping blank videos. So I can't comment on which specific recent topics are hitting and which aren't this past month — and I'd rather flag that than pretend.

What I can comment on: the content mix is 0 Shorts and 30 long-form across the last 30 uploads. In the Japanese tech/AI space in 2026, that's unusual. Channels in the same band have been mixing in shorts for discovery for at least a year, often with strong results. If @tech1018 hasn't tested shorts at all, that's the most obvious gap visible from outside data. Not because shorts are magic — they're not, and they convert subscribers poorly compared to long-form — but because they're currently the cheapest way to put your face in front of new viewers in this niche.

The forward thought: if I were advising this creator, I wouldn't touch the Wednesday/Saturday long-form schedule. The discipline is the moat. I'd add a Friday shorts experiment, three months, twelve uploads — pull the strongest 60-second insight out of each week's long video and post it as a vertical clip. Then look at whether subscriber growth shifts. With 1,400 long videos already in the catalog, there's a near-infinite supply of source material to clip from. That's the kind of advantage a 100-video channel doesn't have, and it costs almost nothing to try.