Notes on using Aside for everything from YouTuber recommendations to coupon checks to tallying up the orders
As soon as Singles’ Day (광군제) kicked off — China’s November 11 shopping festival, which Korean shoppers mostly know through AliExpress deals — YouTube was flooded with recommendation videos. Best-value picks, record-breaking discounts, channel-exclusive codes. They’re tempting, but watch them one after another and you start losing track of what you actually needed in the first place.
This time I tried something a little different. I opened Aside, an AI browser, and started by writing down what I wanted to buy.
A power bank for travel and a tidy charger, a toothbrush sanitizer, a tool to help with cleaning the bathroom, something to help deal with food waste, and a high-power blower for cleaning the car.
Then I asked this:
“Look through at least 10 YouTube videos — don’t watch the videos themselves — and use the specs, prices, and purchase links in the descriptions to put together a list of what’s worth buying this time.” (“유튜브 10개 이상 살펴보고, 직접 영상을 보지 말고, 설명란에 있는 스펙·가격·구매 링크를 참고해 이번에 사면 좋을 것들을 정리해줘.”)
That was the starting point. From there I kept the work going with follow-ups like “Dig up more, including TechMong’s videos” (“테크몽 영상 등 더 찾아봐”), “Find other YouTubers’ coupons too” (“다른 유튜버 쿠폰도 찾아봐”), “Go to AliExpress yourself and check the options and prices” (“직접 알리에 가서 옵션과 가격을 확인해”), and “Leave out what I’ve already bought” (“내가 이미 구매한 것은 빼줘”).
Having done it, I’d say the heart of AI shopping lies less in getting a recommendation list in one shot than in turning search and verification into one continuous task.
1. Not watching YouTube doesn’t mean the YouTubers stopped mattering
The description boxes of Singles’ Day recommendation videos hold more information than you’d expect.
Product names and models, purchase links, which option to select, the listing price, store coupons, channel-exclusive codes, limited quantities, sometimes even the conditions for card discounts.
Rather than asking the AI to summarize the videos, I asked it to gather this information first. It wasn’t a substitute for watching the videos; I had shifted the first step of narrowing down candidates so that it centered on the description boxes.
What mattered here was keeping a record of where each recommendation came from.
An answer like “this charger is good” isn’t enough. I need to be able to trace it back to who introduced it in which video, and what options and codes were written in that description.
For example, these are some of the sources I confirmed in this round of research.
- Balpum Paneun Namja (발품파는남자) Singles’ Day video: the Baseus (베이스어스) 45W mini power bank and related codes
- TechMong (테크몽) Singles’ Day video: the Baseus 100W car charger and store/channel discounts
- HomeLabs (홈랩스) Singles’ Day video: options and codes for the Toocki (투키) retractable cable, the Ulanzi (울란지) card reader, and others
- Dr. Jang’s Shopping Addiction (장박사의 구매중독) video: the Baseus 240W retractable cable and its discount conditions Of course, a description box is not the same thing as independent quality testing. It may include sponsored content or affiliate links, and to know how a product actually feels in use you still need to watch the video or look for separate reviews.
Even so, it was useful as a starting point for deciding what to check and where to begin.
2. Why “find other YouTubers’ coupons too” turned out to be useful
The first recommendation link you find isn’t necessarily the best one. Several channels feature the same product, and each channel is sometimes allotted a different code or a different quantity.
While looking into the Baseus 45W mini 10,000mAh power bank, I asked this:
“Find other YouTubers’ coupons too.” (“다른 유튜버 쿠폰도 찾아줘.”)
That instruction changed the unit of research. It moved from reading one particular video’s recommendation list to cross-checking multiple sources around the same model.
Balpum Paneun Namja’s description listed a store coupon KRVIP1, a channel code BALPUM009, and an expected price of $19.83 for this product. The AI also turned up codes from other channels.
But finding more codes doesn’t mean getting that much more off. Codes from different channels may amount to the same benefit, may already be used up, or may apply only to a specific option.
So I had to keep these two statements apart.
- I found a code in a description.
- I applied it to my order and the amount I actually paid went down. In my order history, the power bank was confirmed at $19.83. But the fact that the expected price in the description matches the order amount doesn’t automatically prove which code took off how much.
Finding a code only gets you a candidate; the verification happens on the checkout screen.
3. What mattered more than the recommendation link was the “selected option”
Partway through the shopping, I also asked something like this:
“Someone recommended a retractable charging cable — which one was it? Link and code. Not TechMong.” (“릴형 충전 케이블 추천받았는데, 뭐였지? 링크와 코드. 테크몽 아님.”)
People remember products less by exact model name than by context — “that cable that reels in, the one so-and-so showed.” An AI browser that keeps the conversation going helped me recover that memory.
The research record let me tell two candidates apart.
The one Dr. Jang introduced was the Baseus 240W retractable cable, and the one HomeLabs introduced was a 3-pack of Toocki 100W retractable cables. The wattage, the bundle, and the codes were all different.
The Toocki product in the HomeLabs description was 3pcs 100W Cable, code HLBS24, expected price $8.20. But what I actually bought was the 5-pack at $10.92.
In a case like this, comparing only the total in the description throws your judgment off. You have to match the pack size and look at the per-unit price alongside how many you actually need.
The same went for chargers. The UGREEN (유그린) charger with a built-in retractable cable had 45W, 65W, and 100W versions and several plug options all mixed together on one page. Finding “the link to the 100W charger” wasn’t the end of it.
What you need to ask of the AI goes beyond handing over a link.
“Check which option is selected right now. Separate the option shown as the lowest price from the option I’m actually trying to buy.” (“지금 선택된 옵션이 무엇인지 확인해줘. 최저가로 표시된 옵션과 내가 사려는 옵션을 구분해줘.”)
4. I entered a coupon, and it actually made things worse
The moment I remember most from this whole experience was the final checkout step for the 1,000-Won Mart (천원마트), the section of AliExpress’s Korean storefront for very cheap bundle-deal items.
The checkout screen showed an item subtotal of $59.80, a discount of $14.48, free shipping, and a total of $45.32. I asked the AI to enter any last code that might apply. I also set a condition: do not complete the payment.
It tried PDEX08, but there was no sign of the final amount going any lower.
Then, when it entered LJMW04, the site popped up a confirmation saying that using this code would take $4 off the discount amount. The total was about to go up to $49.32.
It cancelled the code and checked again that the original discount and the $45.32 total had been restored.
This changed how I think about coupon hunting.
What mattered was less the fact that a discount code exists than how it interacts with the discounts already in place.
The AI’s job can’t end at reporting “code entered,” either. It has to compare the total before and after, cancel if the result is worse, and confirm that everything is back to its original state.
5. Not “a good product” but “where will I use it, and how many?”
The toothbrush sanitizer is a case where the use became more specific as we talked.
At first I just said I needed a toothbrush sanitizer. Later I asked it to look for more portable mini ones, and split the need into the bathroom, the office, and family trips.
This is the final lineup I settled on.
- Bathroom: 1 Hubdic (휴비딕) HTS-20
- Office: 1 Hubdic HTS-850
- Family trips: 4 AiQUE minis Once the needs were split up like this, the criteria for comparing products changed too. For the bathroom I looked at how it holds multiple toothbrushes and whether it has a drying function; for travel I looked at size and price.
With AiQUE, the price and even whether it could ship varied by color. When I checked, the 1,000-Won Mart had green at $4.51 and white at $4.69, and pink was marked as unavailable for delivery. The actual order was two green and two white.
The AI, for its part, suggested trying a small quantity first given the chance of defects, but I decided to buy all four for family trips at once. That kind of choice is the user’s to make. The AI can flag risks and help compare against my purpose, but it can’t set the priorities of my everyday life for me.
Nor did I take the name “sanitizer” or the seller’s “99.99%” line as proof of performance. Reading shopping information and proving that a product works are different things.
6. Why you shouldn’t compare only within AliExpress
Before buying the two Hubdic products, I asked this:
“How does this compare with the lowest domestic prices on Coupang or Naver Store?” (“국내 쿠팡이나 네이버스토어 최저가와 비교하면 어때?”)
When something is part of a sale event, it’s easy to assume AliExpress will be the cheapest. But once you factor in shipping, domestic coupons, and how easy returns are, the answer can change.
My AliExpress order for the two Hubdic products came to 36,576 won. That was favorable compared with the domestic prices I found at the time, but I didn’t go as far as confirming the definitive lowest price across every seller on Naver (네이버), Korea’s main portal and shopping-search platform.
The TOWNEW (토뉴) T3 refill bags went the other way: I didn’t rush to buy them on AliExpress.
The model number of the unit I own is TWN1005. When I checked the official guide, the refills for the T3 were listed separately as TWN2003. On AliExpress, T1 and T3 were mixed together on the same page, so the low price in the search results sometimes turned out to be the T1 option. Selecting T3 sometimes meant it was sold out, or that shipping was added.
The input that helped here wasn’t “find TOWNEW bags” (“토뉴 봉투 찾아줘”) but “TOWNEW T3, model number TWN1005” (“TOWNEW T3, 모델번호 TWN1005”).
When buying compatible parts, the exact identifier the user supplies can be more decisive than the AI’s ability to search.
7. The AI missed things too. So I reconciled the totals
At the end, I asked it to put everything I’d bought that day into a Google Sheet.
Here too, I kept “the purchase-complete screen is open” separate from “the payment actually shows up in the order history.” The card reader first appeared as payment being confirmed, and I later verified that it had changed to processing.
The 1,000-Won Mart purchase was split into orders by seller. The 15 orders in the first pass added up to $40.24. But the total just before payment had been $45.32.
$5.08 was missing.
Checking further under “View more orders” (‘주문 더 보기’) in the order list turned up three more.
- Sonicare (소닉케어)-compatible brush heads: $2.20
- Shoe laundry bag: $1.09
- Pot lid and spoon rest: $1.79 They added up to exactly $5.08. With those included, the 1,000-Won Mart came to 18 orders and $45.32, which matched.
In the end, the orders I confirmed that day numbered 26. I recorded them by currency: $206.97 for the orders shown in dollars and 36,576 won for the orders shown in won. I also organized it so that the total for an order bundling several items wasn’t added again for each item.
This process showed the AI’s limits as well. Assume that the list you see first is the whole thing, and items go missing. After the quick tidy-up, I needed the simple check of counting the orders and adding up the totals.
8. I think I’ll ask this way next time too
Based on this experience, next time I plan to make my requests in this order from the start.
First, write down the intended uses and quantities.
“I need a power bank for travel and toothbrush sanitizers for the bathroom, the office, and family trips. Sort the candidates by my uses.” (“여행용 보조배터리와 욕실·회사·가족 여행용 칫솔 살균기가 필요해. 내 용도에 맞춰 후보를 나눠줘.”)
Second, keep a record of where recommendations come from.
“From the descriptions of at least 10 YouTube videos, collect the product name, exact option, link, price, and coupon conditions. Don’t play the videos, and attach the channel name and the original link.” (“유튜브 영상 10개 이상의 설명란에서 제품명, 정확한 옵션, 링크, 가격, 쿠폰 조건을 모아줘. 영상은 재생하지 말고 채널명과 원문 링크를 붙여줘.”)
Third, look for other codes for the same model.
“Don’t limit it to one YouTuber — find other channels’ codes for the same model too. Keep store coupons and channel codes separate.” (“한 유튜버에 한정하지 말고 같은 모델의 다른 채널 코드도 찾아줘. 스토어 쿠폰과 채널 코드는 구분해줘.”)
Fourth, verify on the actual sales page.
“Check the price and shipping for the option I selected. Compare the total before and after entering a code, and cancel it if it’s worse. Do not pay.” (“내가 선택한 옵션의 가격과 배송비를 확인해줘. 코드 입력 전후 총액을 비교하고, 더 불리하면 취소해줘. 결제는 하지 마.”)
Fifth, check domestic prices and what I haven’t bought yet.
“Compare with domestic prices including shipping. If you can’t confirm it’s cheaper, say so. Leave products I’ve already bought out of the candidates.” (“배송비 포함 국내 판매가와 비교해줘. 더 싼지 확인되지 않으면 그렇게 말해줘. 이미 산 제품은 후보에서 빼줘.”)
Finally, leave a record.
“Go through today’s order history all the way to the end and organize it in a Google Sheet. Include option, quantity, order number, price after discount, and status, and make the totals add up.” (“오늘 주문 내역을 끝까지 확인해 구글 시트로 정리해줘. 옵션·수량·주문번호·할인 후 금액·상태를 넣고 합계도 맞춰줘.”)
The value of AI shopping wasn’t in buying more
Aside was useful this time, and not only because it produced plausible-sounding recommendations. The work carried through as one chain: starting from YouTube descriptions, moving to the actual options and checkout screens on AliExpress, checking the order history, and then leaving a record in a Google Sheet.
That doesn’t mean everything resolved itself automatically. There were products for which it couldn’t find a dedicated coupon, options that were sold out, and items it missed in the order list. I also didn’t measure how much time I actually saved, or the lowest price across every possible seller.
But the roles did become clear.
I set the needs and the budget, corrected course along the way, and made the final call on each purchase. The AI took on finding sources, cross-checking options, testing codes, and checking for omissions.
The best way to use AI on Singles’ Day, I think, is not to stop at asking “what should I buy?”
“Find out who recommended it. Check that it’s the right option. Verify that it’s really cheap on my own checkout screen.” (“누가 추천했는지 찾아줘. 그 옵션이 맞는지 확인해줘. 내 결제창에서도 정말 싼지 검증해줘.”)
When I asked that way, the AI went beyond being a recommendation-list generator and became a fairly capable shopping assistant.
The prices, coupons, stock, and order statuses in this post are as of when I checked on November 1, 2026. The codes mentioned are not guaranteed to still be valid, and this is not a performance review written after receiving and using the products. The linked YouTube descriptions may include sponsorship or affiliate-revenue disclosures.
