It was a real bummer to see our former vet on here. Manchac Animal Clinic is a real place that we took our dogs to for a long time. The docs and staff there are amazing. Seeing their storefront and pictures used like this just… sucks. But as the author says, I guess this is just the world we live in now.
Agree. Also as a PM, one of the most important things you can do, once you've decided or understood "this is the most important thing we can be working on," is to sell that vision internally so you have high alignment and thus can move fast with shipping (i.e. Sales, Marketing, Legal et al are all on board and ready to help you get to market).
SHOWING internal folks something is always more compelling than words on a slide. So if you have good alignment with your Engineering team that the thing is possible, you can go off on your own and vibe-code something to garner internal alignment... and then throw it away and let your Engineers build the real thing. And they won't have to spend extra time on a PoC that's only meant to show off internally.
Those vendors likely won’t accept hardsubs, because it would mean 10 video files for 10 languages, instead of soft subs where you get 1 video file 10 languages (10 different subtitle files).
But here’s the other thing - CR could have used the ASS subs on their website and given the less-dynamic sub files to their vendors. You can save a master subtitle file in whatever format you want.
Well, they can, and at least did. I know because I was one of them! The P&L that I rolled up to our execs was dead simple as well. I think everyone had a pretty clear picture of what was going on, down to the fraction of the hour.
It’s cheaper than you might think. Much like in gaming, there’s a lot of people who really want to work in the anime industry, even if it’s just on the localization and distribution side. This drives down salaries quite a bit.
It’s a great question actually, and the answer has mostly to do with Romanization. Japanese and English are sufficiently structurally different, that even the sentence length won’t necessarily be one to one (eg subject and object inverted).
Another thing that happens is time code shifts that come from differences in frame rate between source material and what the subtitlers end up with (eg 24 vs 23.98 if I’m remembering correctly), which can cause subs to have what we called “ramping” issues over time (timing gets less and less accurate). So you have to go through and reset all the lines anyway.
That being said, we DID do this sometimes, but maybe that takes your time down to 25 minutes, the hard minimum possible time to accurately subtitle a 25 minute show.
And translators hated having to add the times codes (or copy paste their translations over the CCs) — they preferred to just give a script to the subtitler and let them handle it. And actually, if it’s a really good subtitler, they can! In about 35 mins.
So I think the translators were probably right to push back, as it’s only 10 minute savings for probably >10 mins on their part.
Hah, this is my time to shine. I worked in anime subtitling and timing for a number of years. I helped write our style guide — things like how to handle signs, overlapping dialogue, colors etc.
It wound up being quite a large document!
But the thing to realize here is that, all of these subs have to be placed by hand. There are AI tools that can help you match in and out times, but they have a difficult time matching English subs to Japanese dialogue. So what you have to do is have a human with some small grasp of Japanese place each of these in/out times by hand.
If you’re really good you can do one 25 minute episode in about 35 minutes. But that’s ONLY if you don’t spend any extra time coloring and moving the subs around the screen (as you would song and sign captions).
Elite tier subs can take up to two or even three or four hours per episode. That’s why the best subs, are always fan subs! Because a business will never put in 8x more time on an episodes subtitles than “bare minimum.”
Crunchy roll looks to have at least gone halfway for a while… but multiply those times across thousands of episodes over X years… and you can see why some manager somewhere finally decided 35 minutes was good enough.
I am in the Product world now, and I do think this was a bad move. Anime fans LOVE anime. The level of customer delight (and hate) in the anime industry is like no other. I really miss the excitement that my customers would get (and happily telegraph!) when I launched a product in those days. Which is all to say, you HAVE to factor delight into your product. Especially with a super fan base like you have in anime.
From a purely supply chain perspective, how should we be navigating this as a hardware startup?
I also read today that a NATO official "urged businesses to be prepared for a wartime scenario and adjust their production and distribution lines accordingly."
How should we be thinking about the next 4 years? Sourcing from US suppliers exclusively?
We've been lucky I guess that up to this point we just had to think about the tech. But now the supply chain for our product is suddenly as complex a problem set...
I'm just... Incredibly suspicious of this. I've just this week received an email from Klarna that "effective October 10th, 2024, JPMorgan Chase will no longer allow customers to make Klarna payments using their JPMorgan Chase credit cards."
Surely this can't be a coincidence. Two points I think: one, it just doesn't pass the smell test that you can cut half your staff and magically become more efficient (yes, Twitter, but they're not doing so hot from a valuation perspective)... Especially in FinTech?? And two, it really seems like they're using "AI" hype to obfuscate this fact from folks that don't really understand it's practical implications, in the face of a spiralling business.
That's fair enough I think! I totally agree that larger backlogs tend toward staleness. Just like a good shrub, they need care, feeding, and trimming :)
By the way, I enjoyed the article! It's a good share, and appreciate the perspective.
I appreciate the sentiment here, but this is not my experience. I talk to our customers almost every day, and for this reason, we have all kinds of well-vetted features and improvements in our backlog. There's always more to do than you have time for; the difference is whether or not your backlog is directly informed and refined by customer feedback.
Yep I work on Search as well -- both of these books are excellent.
Website search is... hard. A lot of the faceting still needs to be done by hand. I think there's probably some opportunity for LLMs to make some sort of autotagging/categorization easier, but there will likely still need to be a human in the loop to verify.
What about like, blogs, for instance, as a source of training data? Is the differentiating factor here something like, it's lots of people interacting with other people?
Yes I did this. It was pretty tough but there is a way.
I was fired from a job in the media industry and decided to start over in tech. It was really a low point for me honestly; I can empathize I think with your post.
I found an entry-level job as a Scrum Master. The role gets a ton of hate, and some of the criticism is justified, but it helped me learn the tech world from the very bottom.
I spent about 4 years doing that. I did a lot of listening and learned an incredible amount from the development team. It was probably the most important thing for me -- whatever career you choose, find a way to start in the trenches with the folks doing the real work.
Then I became a Program Manager. That was really good because it helped me learn the delivery aspect, and get me plugged into the strategy aspect of the techbiz.
I should note here by this time that I had been really lucky in a couple of respects:
1. I found a company that was VERY good at developing their employees into new roles.
2. I found mentors and sponsors within that org that helped me to grow and sponsored me for new roles.
But after a few years learning the Program side, I transitioned to PM, where I am very, very, happy, and to be honest have a leg up on some of my peers because of my deep experience embedded with dev teams and understanding their struggles day to day, but also having to service the needs of the leadership/exec team as a PgM.
I guess I would sum this up by saying, it sucks, and it takes time, but it is 100% doable. I'm not exceptional in terms of brains or skills, which I don't mean as a knock to anything or anyone, but rather as a way of saying, "If I can do it, I really do believe others can too."
The thing to focus on are my 1 & 2 points above. Those are absolute game changes and IMO must-haves.
And LUCK again plays an unfortunately large part in all of this, I feel I must repeat. But I wish you the best of luck, and hope that my note gives you an amount of hope.
I'm a huge fan of contemporary poetry. I can give a few recommendations, but of course you'll have your own tastes, so I'll try to annotate :)
"Here, Bullet" by Brian Turner. Poetry of the Iraq war.
"In the Surgical Theatre" by Dana Levin. Medical, experimental.
"Black Aperture" by Matt Rasmussen. About his brother's suicide.
"Homie" by Danez Smith. Hard to pin down but amazing book. Stream of consciousness, experimental, explores queerness and blackness.
"The Tradition" by Jericho Brown. Phenomenal. Explores being black in America, more historicism than Danez's collection.
"Frank" by Dianne Seuss. Powerful and beautiful. Modern lyrics. All sonnets. Will change the way you think about what poetry is. Stunning collection.
"Post colonial Love Poem" by Natalie Diaz. What it sounds like :) really cool book
"When I Grow Up I Want to be a List of Further Possibilities" by Chen Chen. A little bit more fun.
All of these, of you're not familiar with contemporary verse, will challenge your notions of "what poetry is."
Here's my advice: just read it and experience it. It's supposed to be fun! None of these works are puzzles to be solved, or have hidden secret meanings. They are all works of art, and all experiences of them are valid. Enjoy!
This article is about organizations, but I'm curious how this plays out on web properties as well. The larger ones I imagine make use of many different data sources, and that data needs to be federated somehow, or one loses out on the insights from the aggregated data.
Does anyone have any experience for how this is well architected?