To HN users reading this comment, where did you find the best reader view implementation?
Seems like it's a really difficult task.
322 karma · joined November 8, 2019
Seems like it's a really difficult task.
1. Data about what the population is interested: More people use Google today, giving Google more data to train its AI with.
2. Server capacity: It's cheaper for Google to store 1 image than for me to store 1 image, because it stores many more images. Same for compute, such as training AI. Google can get one A100 GPU cheaper than I can.
3. Ads: If I had the exact same search quality and even the exact same number of users, Google will make more money from the same operation, because it has more advertisers.
They set up their business well!
Thanks for helping me see this!
I'm not embodying any fake persona, just trying to find what I would find interesting. I just don't get addicting content. I even try searching so that TikTok would get explicit keywords from me.
I subscribe to a bunch of email newsletters and RSS feeds, check Reddit and Twitter daily, and check HackerNews morning and night, and I have no problem finding interesting content every time I open these things (Twitter being least interesting to me out of this list).
I like software engineering, business, product management, marketing, and technology industry generally. I don't believe I am so far from regular consumers: I check out mainstream videos on YouTube like MKBHD and even subscribe to his podcast (This is to show I find content for "mainstream consumers" interesting, though I work in tech and know more about tech things than MKBHD's target audience).
Can someone help? Why is the whole world praising TikTok while I am failing repeatedly?
A teammate once told me he is frustrated with being asked to work on an MVP, that is minimum cost, minimum polish, minimally functioning. He said he wanted to work on MLP, so I said he can call it MLP if he wants, but what we must build does not change: prove the concept minimally such that we can verify whether the problem we seek to solve is a real problem in the first place, then we can add as much polish as needed once we know it's a good problem to solve. If we think customers are experiencing some problem that, in order to solve minimally, we need to give them polish, then the polish is a part of the MVP/MLP. Otherwise, it is not. It doesn't matter what we call it. What matters is what we must do.
He thought that my attitude taking the distinction between MLP and MVP lightly was a sign that I do not care about the best interest of our customers, and left the company.
If you mean waiting idly wishing that a particular event would happen in this world, then no.
If you mean taking the same event that happened, and finding how to find ways to make the most out of it, then yes.
It is always about luck, not one you wish to receive from the outside, but one you create from the inside.
A spends time and effort thinking about how to block B from getting better.
B spends time and effort thinking about how to be better for customers.
Which do you think will win? Or should win?
Guide your actions accordingly.
If MVP does not solve a real problem, it does not validate the correct hypothesis, so what the post describes as an MVP that is not MLP is not an MVP.
And an MLP executed to be not MVP would be a thing built to be built, or taking unnecessarily risky bets spending unnecessary cost.
Cut the crap with fighting words with words. If the community used the word MVP to help introduce a more useful concept, real builders focus on executing it correctly, to the best of our ability, giving the inventors fair credit and most charitable interpretation. Alas, we have content marketers among us, and those who failed and wish to blame it on being given wrong instructions to follow, inventing empty words like MLP.
A dad joke bot!
But I think this simple example suggests conversational forms can be unintuitive for consumers, and it can be more costly for both the asking side and the answering side to communicate the same information.
What was this part of experience like?
When you made this, why did you believe building this was what people needed?
And what did you end up observing that made you discontinue the project?
In former, hypothesis specifies which variables are likely independent, dependent, must be controlled, so on, then makes observations to confirm. In latter, data is first collected with as many variables as possible, then we attempt to identify which variables are independent, etc, then confirm.
To be cost efficient, many science experiments chose former. AI models use latter (get data, then find patterns, aka dependencies across variables). But even in latter, 1) human scientists can and do use the approach when appropriate, and 2) someone must have chosen the variables to observe which is a hypothesis (but lower precision is permitted in favor of higher recall).
Long story short: It is sensational and in the best interest of the media to make it seem like some drastic change is happening, but the way we the humankind is growing our collective knowledge is still fundamentally consistent, except with some added techniques thanks to low cost of studying large scale data, and it would serve us well to focus on the fundamentals and basics. The scientific approach.
What matters is not who predicts which future, but rather who creates which future.