Mermaid: Markdown-like generation of diagrams and flowcharts from text
57 tomcam 3 hrs 13
Unable to load the content https://github.com/knsv/mermaid
- My Posts
- Chat to Authors
Home
Home
Member since 2017-07-15T03:50:57Z. Last seen 2025-01-12T00:12:02Z.
2732 blog posts. 128 comments.
57 tomcam 3 hrs 13
Unable to load the content https://github.com/knsv/mermaid
58 edouard-harris 11 hrs 60
Hi HN! We're Ed and Jeremie, the founders of SharpestMinds in YC's W18 batch. We're building a free online community for ML/AI developers through which they can access job opportunities. (You can apply to join it at https://www.sharpestminds.com/members) We're ML developers from non traditional backgrounds. Ed did a PhD in biological physics, and Jeremie studied quantum optics before dropping out of grad school to work on SharpestMinds. We started looking for ML jobs after school, thinking it shouldn't be too hard to get one. We found to our naive surprise that we fell short on a number of skills that are needed to do good work in industry. You just don't learn much devops in grad school.
As a result we decided to build something that would make it easier for ML devs to develop (and discover!) skills they might be missing, and then get their first jobs or internships. From the outset we also wanted to build a community around the process, since looking for your first job is usually a pretty lonely experience. Because we monetize directly through hiring, we can afford to create a space for discussion without ads or algorithmic distractions :)
Our typical users so far have been grad students who know ML material well, but don't yet have much, or any, practical experience. However, you don't need a degree at all (a few of our users are self-taught high school dropouts), and anyone who knows the material is welcome. In fact, that's one of the advantages of our system: we test directly for knowledge, so it doesn't matter how you got that knowledge or how long it took you to get it. One of our goals is that by the time we present you as a candidate, things that would otherwise be holes in your resumé don't matter so much, and we can make that case to companies that are hiring.
To qualify for joining, you do an online deep learning quiz (here: https://www.sharpestminds.com/members/apply), followed by a technical interview. If you pass both, we invite you aboard. It's possible to retake the quiz a month later if you don't pass it, and we'll send you tips on what to study in the meantime.
Once you join you get access to a job board with exclusive (i.e., not scraped) internship and full-time opportunities on it. We've created an application system where your profile gets customized to the job you're applying for, to maximize the odds that you'll get an interview. We also have lists of common interview questions, mentors that you can practice interviewing with, and periodic AMAs with ML hiring managers from companies like Skydio and Airbnb.
The hardest part about building this has been figuring out the best way to present our users to employers. Early on we found that hiring managers were passing on qualified people, because their eyes would glaze over from reading too many CVs. We ended up building application profiles that let our users display their most relevant personal projects prominently in their application. The interview rate has increased significantly as a result.
If our approach works for the ML/AI field, we'd like to build communities like this for other fields too.
We're looking forward to getting feedback and hearing ideas from HN! We know there are lots of ML devs / enthusiasts on here, and we'd also be very interested in hearing about your own experiences making the transition, or similar programs you might know about. We'd also be interested in hearing about what, in your experience, are the most important programming skills needed by someone with a good knowledge base but little practical experience to be a strong contributor at their first job or internship.
esja 9 hrs I recruit in this area and wish you all the best. One piece of advice: change your name. In my experience the sharpest minds are intellectually honest to a fault, and would rarely describe themselves that way.
exolymph 8 hrs I'd go further and say: The current name is outright cringeworthy.
DoofusOfDeath 5 hrs Agreed. The name seems like an advertisement that you think everyone outside your company or community is less intelligent. IMO that crosses the line separating puffery / self-promotion from outright insult.
edouard-harris 3 hrs Definitely not intended that way. We'll put some thought into changing it based on everyone's feedback.
edouard-harris 9 hrs Thanks for the feedback! We're open to a change. Let me know if you have any thoughts.
BryBran 7 hrs The Neural Network
edouard-harris 7 hrs Solid
harigov 4 hrs neural.network
ShannonAlther 7 hrs Seconded.
dang 6 hrs That is seriously good.
seshagiric 8 hrs how about "the ai folks". Emphasis on 'the'
auvi 7 hrs or as Sean Parker would say: Drop the "The", just "ai folks".
cproctor 7 hrs cogs
p0d 9 hrs Funny, had the very same thought.
ianbicking 1 hr A couple bits of feedback:
At least in the title of this post it is called an "Online Community" but it really feels like a job board. Is it really a community at all? I expect to feel disappointed.
Wait... is it a job board, or an internship matching system?
I got about 5 duplicate questions.
The timed questions with a big code block and multiple choice were stressful, in that there was some dense code to read and I couldn't decide whether to understand the code first or read the questions first.
It wasn't clear to me that the timer was actually a limit, and not just a suggestion (i.e., something to pace yourself to do all the questions in the time limit)
The SQL questions felt like very normal SQL questions. They seemed easy enough (assuming I got them right!) simply given past experience with database driven websites.
edouard-harris 39 mins Thanks for the feedback here. Much appreciated.
1 & 2. It's.. both. From the inside it feels like Slack + job board + GitHub-like profiles
4-6. Noted, we'll keep updating the UI & question bank. Knowing this about how the timer feels is especially useful
minimaxir 10 hrs
However, you don't need a degree at all (a few of our users are self-taught high school dropouts), and anyone who knows the material is welcome.
To qualify for joining, you do an online deep learning quiz (here: https://www.sharpestminds.com/members/apply), followed by a technical interview.
Once you join you get access to a job board with exclusive (i.e., not scraped) internship and full-time opportunities on it.
These three constraints don't reconcile with each other.
Yes, new ML/AI resources like TensorFlow and MOOCs have made AI more accessible, and that having a degree is no longer required to implement ML/AI. I agree it's unnecessary gatekeeping to require a degree to be able to play with ML/AI.
But what showy YouTube videos and Medium thought pieces don't teach is implementing ML/AI in practice to solve business problems. The stereotypical quiz + technical interview for the ability to join the service won't account for that.
When I was looking for jobs last year, 100% of the job openings for ML/AI (as opposed to Data Analyst/Data Scientist) required a Masters/PhD. In that case, I can't blame them, since there is a certain amount of experience and knowledge required to define problems and work up statistically sound solutions that can't be done by simply adding layers to a neural network or ensembling XGBoost models.
edouard-harris 10 hrs You're right, but this is the same catch-22 for getting started in any nontrivial field. How do I get experience if getting a job requires experience?
Having an MSc / PhD in the field doesn't resolve this. HR departments use grad degrees as first-pass filters, and thereby miss self-taught people who are genuinely competent.
We try to solve for this by easing people into jobs with internships and work terms first. The community is a key part of that since it supports them if they get stuck on an implementation problem. And of course we're incentivized to make sure members perform well in the internship phase, since we make money when they're hired full time.
ujal 9 hrs I completely agree. Machine Learning is on the way to become a field like Web Development. There is a huge supply/demand gap that will only get wider.
DoofusOfDeath 5 hrs With the caveat that holding a PhD is positively correlated with success in ML, and negatively correlated with success in web dev.
(I'm guessing.)
wakkaflokka 10 hrs I always felt it was the other way around - using MOOCs or online tutorials makes applying machine learning to business problems in practice a whole lot easier. But in terms of inventing new algorithms, deeply understanding the theory and origin of things, and doing active ML/AI research, that still seems to be in the realm of requiring a PhD (i.e. the difference between a "data scientist" and an ML/AI researcher).
minimaxir 10 hrs That's true: the reason I started looking into AI is to broaden my knowledge on how it can be used to solve problems in ways traditional ML/traditional modeling can't. I did have a strong statistical/data background in college beforehand which helped validate which approaches were good/bad/wrong, though.
The YouTube/Medium posts however advocate "Learn Machine Learning from scratch in 3 Months!" which is a problem.
aeorgnoieang 6 hrs How long does it take to 'learn machine learning'? I'm pretty sure I know some about it. I can certainly apply some basic algorithms and there's no shortage of info about lots of variations on the basics I do already know. So what am I missing? How much would I have to learn to be able to claim, per your unstated standard, that I've 'learned machine learning'?
donovanr 9 hrs Some feedback on the quiz:
a few of the questions were very good, and either spoke to key high level concepts, or were specific while being language agnostic. (e.g which one of these layers wouldn't you need, why wouldn't this type of classifier work on this data).
too many of the questions were hyper-focused on the minutiae of word embeddings, tensor flow syntax, SQL queries, and recommender schemes.
many of the questions were constructed vaguely enough that "I don't know" would be the technically correct answer even though I don't think that was what you were going for.
metadata: recent PhD with serious grad courses in ML and working in DL/CV for the past year using a non-tensorflow framework (PyTorch).
edouard-harris 9 hrs This is great feedback. Thanks!
We're constantly iterating on the quiz and it would be great to get more detailed thoughts on it.
If you'd like to do that, please get in touch! (Email in my profile)
donovanr 6 hrs would be happy to, but I don't see your email there -- mine's in my profile (I think!) if you'd like to get in touch
edouard-harris 6 hrs Sorry, realized it wasn't public. Just updated, should be there now!
ellisv 6 hrs Took the quiz and completely agree. Most questions were either overly concerned with detail or too vague.
High-level I don't think a quiz is necessarily the right tool either. Reminds me too much of taking the SAT or GRE.
edouard-harris 5 hrs Yeah we definitely aren't convinced that a quiz is the optimal format for this evaluation.
Statistically, it does an OK job at being an initial filter. My biggest concern at the moment is that it's too coarse of a tool and it might be mistakenly turning away competent people.
Definitely a work in progress. If you have ideas on alternative formats or better questions, please email me. (Email in my profile.)
TekMol 9 hrs First thing my sharp mind noticed is that the page sends data to connect.facebook.net
Why does Facebook have to know what I am doing jobwise?
edouard-harris 8 hrs Good catch, this is a leftover facebook pixel from back when we were experimenting with FB ads. Just created a GitHub issue to rip it out
sharemywin 10 hrs Why not create a tiered community. Beginner. Learning. Expert.
Beginner - Anyone with an interest similar HN. Maybe resource to get into the Learning Area.
Learning - Place to find others learning the material. Maybe find other people study with. or collaborate/reproduce projects.
Experts - people actively looking for employment(what you already have planned).
edouard-harris 10 hrs Great idea! We're actually in the process of doing this. Pilot version is 1 week away.
We're starting with 2 tiers instead of 3, but the goal is similar.
dannytatom 10 hrs Not to sound rude or anything, but the name is kinda douchey and outputting.
DoofusOfDeath 5 hrs SharpestMindsAmongstAllNonDouchebags.com ?
edouard-harris 9 hrs Fair enough. Any suggestions?
dannytatom 9 hrs Not a clue, I'm bad at names lol. Sharp Minds would probably be fine, just Sharpest Minds makes it sound like you're talking down to anyone who isn't involved in AI (which is like 99.99999999% of people).
Or if it was a clever reference to something in AI that made you sound smart but because it was a clever reference comes off more funny than rude.
dhimes 7 hrs I actually took it as people who were building the sharpest minds, not necessarily those who had them.
ImSkeptical 9 hrs How about "Mind Makers". I think many people will have a natural resistance to joining a community named after how smart the members are. Imagine wearing a button that says "I'm the smartest".
mkagenius 7 hrs Advice: Hold kaggle like contests -- ML/AI developers/enthusiasts come flocking to these. When you are at it, please don't make it suck like Kaggle.
nlowell 8 hrs I have been out of college and working fulltime in software for two years. I have been spending a large amount of my free time learning DL through fastai. I really want to move to a job that involves deep learning, and I would like to use your system, but I don't feel like I can responsibly switch to part time or internship work given the stability and pay of my current job. Do you have any option for people who want to go direct to fulltime?
edouard-harris 8 hrs Yes.
You can either (1) keep going with your current job and do this on the side assuming your employment contract allows it, or (2) apply to a company that wants to do full time right away without a work term.
There may be other options depending on your situation, so email me if you want to discuss further. Email address in my profile.
JJseiko 10 hrs I like the idea! Very much actually. It's a real problem - the discrepancy between actual skills and how your CV looks.
In that it reminded me of the way that Basecamp hires - which is that in the "final rounds" they actually hire the candidates to do some small project that is actually needed at Basecamp - just another way of getting at the bottom of what a person can actually do rather than how they look on paper.
edouard-harris 10 hrs You're right, this approach grew out of how we hire people ourselves. Give them a chance to prove themselves on a real project.
Wasn't aware that Basecamp did this too. Thanks for sharing!
ghostbust555 9 hrs Very interested in this, however I would prefer if it was more than just a jobs board. Something like github+reddit in addition to a jobs board would make this very cool (haven't tried it yet so this may already be the case and I am wrong!).
edouard-harris 9 hrs It's more like Slack + jobs board + GitHub-like profiles.
gzeus 9 hrs I had made a job board for machine learning jobs as well https://mljobslist.com/
lexalizer 9 hrs I may have misread this, but would OP agree that ML/AI is not composed of just DL. Currently, the quiz is perhaps too focused.
edouard-harris 9 hrs Would agree.
Yep, quiz might still be a bit too focused. Trying to strike a balance between breadth / depth / time spent answering questions, but still some fine tuning to do
danvoell 9 hrs I'm sure I am making an incorrect inference, but I'm probably not alone. Average work term payment = $5K, average work term = 8 weeks = Below minimum wage.
edouard-harris 9 hrs Thanks for flagging! Embarrassingly I reread this stat and realized that (1) it's way out of date and the true number is higher, and (2) it refers to part-time work (20 hrs / week) but we don't actually say that anywhere.
We've never had a work term work out to less than $25 / hour, except possibly in geographies like India where cost of living in USD is very low. This is below market, but it's only for the work term before getting hired.
Will update this with the most recent stats and clarify part-time status. Thanks again for catching this!
ai_ia 10 hrs This is great. Is it US only or available in India too?
edouard-harris 10 hrs Available in India. We already have several users from there.
ai_ia 10 hrs Awesome.
fwdpropaganda 9 hrs This website doesn't work on my computer.
edouard-harris 9 hrs Thanks for flagging. Do you mind sharing your OS / browser versions?
fwdpropaganda 9 hrs Ubuntu / Firefox
My Firefox doesn't run JS though, so there's that.
ggg9990 1 hr Mine isn’t working either. I also use Ubuntu / Firefox and also the computer isn’t plugged in, so there’s that.
sooheon 8 hrs OK, you can probably guess the reason
fwdpropaganda 8 hrs Yes. Inappropriate technology decisions.
全球最大串流影片公司Netflix，周二公布強勁季績，帶動股價上漲10%並創下歷史新高。其中營業額按年大增四成，盈利更飆六成，不讓人意外。新增用戶達到750萬，也不出奇，真正讓人意外的，是Netflix期內加價14%。
過去10年消費者已給寵壞，認為網上娛樂是不用付費的。市場過去曾擔心Netflix在巨額內容製作到底能否回本，但現在消費者願意付費，說明人們其實是願意為高質素影片付錢的。這不單與Netflix有關，而是整個行業估值提升。
影片行業估值提升
媒體公司面向大眾，卻是向背後的廣告收費，能向用戶收費的企業太少了，所以串流電視出現後，中美科網巨頭紛紛入場。但投資者看到巨額投資，擔心收不回來，故一直只肯給予低估值。但現在Netflix能加價，說明影片生意可為，相關企業的估值看來將出現顯著提升了。至於Netflix，限於其負債深重和經營現金流長期流出，本欄雖不敢再看淡，但也不敢輕易看好。
長影片行業有改善，那麼短片呢？內地短片市場上周出現劇烈變化，新晉媒體「今日頭條」遇上政策收緊，旗下五大手機App中，今日頭條App被勒令下架三周；內涵段子永久關閉；音樂頻道抖音，則關閉直播和評論功能；火山小視頻被下令整改。
今日頭條是2012年才成立，懂得用人工智能推測讀者喜好，送上最合口味的文章，一舉成為全國最大新聞網站。旗下的抖音，則是全國頭三大的音樂短視頻App。市場一度估計今日頭條的估值達200億美元。
內地短片王者估領袖
內地短片市場已經歷過幾個周期，秒拍、美拍、小咖秀等曾一度興起，但最後由快手跑出。抖音曾一度給予強大壓力，但現在看來要先放慢腳步。
短片市場較長影片市場更有趣，因為人的時間和專注力有限，長影片只能回家或乘車時看，短片則任何時候都能觀看。你可能一周都不看一套長影片，但每天「碌」手機時總會看幾套短片。且長影片你會很快按「略過廣告」鍵，但現在不少軟性廣告短片，已經拍得很好笑。
美國有Instagram，中國當然也會有自己的短片王者。抖音放慢腳步了，雖然是用戶的大不幸，但卻是其他短片平台的大幸，投資者可想想誰有機會成為中國的Instagram。
留意直播平台政策風險
至於直播平台，國人口味真的與別不同。女孩子在網上直播唱歌，可以得到上百萬元的打賞。東北大叔表演用橡皮筋炸西瓜，竟然可以賺到買車買房，讓直播平台企業估值提升。
這樣的直播雖不能說是壞事，但感覺總好像有點怪怪的。抖音和內涵段子被整改，便是因為被指內容低俗、色情，甚至有違法情況發生。直播是一種新興媒體，除了要小心用戶口味變化外，也要注意政策風險。
總的來說，長影片的估值有望提升。短片則是美國已有王者Instagram，內地的王者則正在孕育中。直播則模式較新，或許還需要多些時間觀察。但不管如何，影片都是科網裏不能錯過的分類。
hcl.hkej@gmail.com
(編者按：郝承林著作《致富新世代2──科網君臨天下》現已發售)
歡迎訂購：實體書、電子書
A 36-year-old who learned to invest like Warren Buffett explains how saving can actually cost you money If you're just saving and not investing, you're setting yourself up to lose money in the long run. That's a lesson Danielle Town, author of "Invested: How Warren Buffett and Charlie Munger Taught Me to Master My Mind, My Emotions, and My Money (with a Little Help from My Dad)," learned the hard way. When Town, then 34, found herself burnt out as a corporate attorney, she started brainstorming ways to retire faster. "I started to think, 'What else can I do to support myself without being dependent on my salary?'" she tells CNBC Make It. Her first instinct was to hoard as much cash as she possibly could.
"What I was doing was saving money, which I thought was genius," she says. "And I felt very comfortable with my money, figuratively, under my mattress, just protected, and careful, and safe." As she writes in her book, "it wouldn't grow much, I knew, but it wouldn't get smaller either." PLAY VIDEO Warren Buffett, chairman of Berkshire Hathaway Inc. (left), plays pingpong with Bill Gates, chairman and founder of Microsoft Corp. and a Berkshire Hathaway Inc. director, at the Berkshire Hathaway annual shareholder meeting in Omaha, Nebraska, on May 3, 2015. But after talking with her father, author and investor Phil Town, she realized that keeping money long-term in a place where it wasn't growing would leave her with less in the end, thanks to rising inflation rates. "Now, I realize that to some people who know about financial stuff, this sounds ridiculous," she says. "But I didn't know anything about financial stuff. I knew inflation was a thing that felt very macroeconomic, but I had never connected it to my actual savings." In reality, she found, she was "losing money through doing nothing." What happens is that inflation causes prices to rise, which makes money less powerful over time. While a $20 bill will always be worth $20, what you're able to buy for that amount dwindles. PLAY VIDEO Learn Warren Buffett’s simple psychological trick to being persuasive If you had stuffed $1,000 in cash under your mattress 50 years ago, today it would have the same buying power as only $137.45 did in 1968. However, that same amount invested with compound interest would have grown to about $20,000, assuming a 6 percent rate of return. Even if you only earn a 4 percent rate of return, it still grows to around $7,000. Dedicating a solid chunk of money to savings for the future should be a key part of anyone's financial plan, but that money shouldn't sit around gathering dust. Town's father breaks it down in the book: The people who encourage saving mean that you should save instead of spending money. But because of inflation and the fact that what people buy will be more expensive 'tomorrow,' people spend all of their money and whatever they can borrow 'today,' so we need to be encouraged to save instead of spend. Nobody is saying, once you've saved some money, to keep it sitting in a savings account. Rather, it's smarter to put that cash to work. "The antidote to losing money on inflation is investing," says Town, now 36. "You've got to do something with your money." Town chose to make investing a daily habit and researched specific companies to put her money in, but there are plenty of alternatives. Here are a few simple, low-stress ways to start investing: Sign up for your employer's 401(k) plan and take full advantage of any company match, which essentially gives you free money. Contribute to a Roth IRA or traditional IRA, which are both individual retirement accounts that offers tax breaks. Use micro-investing apps such as Acorns, which help you begin by investing small amounts of your "spare change." The app rounds up your purchases to the nearest dollar and automatically put your coins to work. Try other apps that aim to make investing simple and accessible. Consider automated investing services known as robo-advisors that can help you out no matter how much you have in the bank. Research low-cost index funds, which Warren Buffett recommends. Don't miss: Here's how much you'd have if you'd invested the money you lost to the pay gap Like this story? Like CNBC Make It on Facebook! PLAY VIDEO
34 jadeydi 3 hrs 27 news.ycombinator.com/item?id=16858053
If it works, which method are you using now? orliesaurus 42 mins I've been doing SEO as a side gig for many smaller local websites. The amount of traffic you can generate from locally ranking higher than your competitors is astoundingly high. You can literally bring tens of thousands of dollars of extra business to companies i.e. electricians, vet doctors, dentists, gardeners... honestly it's worth every penny.
dx034 13 mins Any chance you could share some ideas? Is it just getting it linked on other pages? That appears to be what a lot of "consultants" online suggest. Or purely content optimisation?
orliesaurus 6 mins A comment wouldn't be enough, and ideas aren't worth a penny without a detailed execution plan. Honestly if you are interested you have to put the hours and do your own research, then test things yourself and see what you are most comfortable doing/what works for your niche or business. If you're really interested, lets exchange emails or telegram and chat, I don't mind helping you out (for free ofc).
chrischen 1 min Do you mostly do content generation, link building, or on-site html optimization.
petra 32 mins How do you do it on the side? Isn't it a very dynamic, very competitive field, so it requires a lot to get good results?
amelius 13 mins I guess it takes a lot of time. Change a thing, wait for Google to crawl your site, then inspect the results and repeat. Most of it is waiting for Google. So yes, I suppose you can do it on the side.
3pt14159 33 mins Yes it does, but a little goes a looooong way to Google.
I've also been starting to notice that having custom HTML / CSS is boosting things more than before for smaller sites.
Also, even if using JS for SPAs doesn't matter to Google there is a reasonable long tail of other web crawlers that may or may not be helpful to your ends and you'd be surprised at how many of them just use mechanize or nokogiri instead of something more robust. Basically it comes down to cost, last time I was in the custom crawler trade it was around 50x more expensive to do things with a virtual dom than it was to either straight parse the HTML or figure out the main API calls JS was making from the browser that actually had the data I needed and just run those.
devin 21 mins Out of curiosity, what options beyond nokogiri, mechanize, etc. are you referring to? Something open source, commercial, or custom?
dmtroyer 21 mins Can someone post or link to strategies for optimizing for local google seo?
jimmy1 2 hrs Absolutely -- for paid search. Organic SEO is a little tougher. You have to really provide valuable, quality content for organic SEO. If you try to half-ass it at all, it will be snuffed out and mixed in with the other garbage. Provide incentive for users to browse your content and don't make it obvious that you are just having a cheap marketing page with a couple of blurbs any idiot can google to figure out the information you are providing. I think that is why sites like Bankrate and NerdWallet are good at organic SEO -- they have nice tools and good information.
Azkar 55 mins
You have to really provide valuable, quality content for organic SEO
This is the key. Google wants to make sure your readers are getting their questions answered. SEO is constantly changing, but high quality and valuable content will always be king.
ilhicas 2 hrs I believe the best methods to remain the same that it has always been, create quality content by delivering value to your visitors and keep your webapp running without maing your page hang on clients as well as presenting a clean interface, that should be cross device.
amelius 37 mins That must explain why w3schools is always at the top of my search results.
/s
openIce 26 mins What's wrong with w3schools?
alexmorenodev 20 mins https://www.w3fools.com/
vinylkey 6 mins
Today, W3Schools has largely resolved these issues and addressed the majority of the undersigned developers' concerns.
Seems to me like it's a decent resource nowadays.
hknd 19 mins it is a shitty site which shows out of date shit.
hashsalt 27 mins Depends on the market. Google's move to 'entities' over keywords means broad search seo is beyond the abilities of "2 guys with some adwords experience" type agencies that were popular circa 2011/12. Quality well thought out sites that (theoretically) align with intent fare better.
That said, if you're targeting a certain city, definitely still works. Most local sites are poorly optimized.
arnon 10 mins SEO works but actual focused content is better.
always_good 56 mins As long as there is search, there are ways to appear towards the top of it.
deadcoder0904 2 hrs Yes. For SEO related content, I always check Backlinko - https://backlinko.com
The guy Brian Dean has many years of experience about SEO
orliesaurus 39 mins Funnily enough reading studf from this site's (i.e. rank brain 2018 guide) on mobile makes my phone lag hard (octa core from last summer). I think the 5000+ heavy images content needs some mobile optimizing... Here's a little audit with Chrome's own tools: https://i.snag.gy/eX7IB1.jpg
kristianc 2 hrs It's probably less relevant than it was, due to Google's aggressive internal linking. Google now surfaces more and more information (sports scores, directions etc) without you having to leave Google. If you run a service competing with Google in one of these areas - good luck.
Similarly, AMP has kind of changed the game in terms of publisher traffic. Users seem to be assuming that results that do not appear in those carousels are more outdated / less relevant.
gscott 48 mins The new SEO is really branding just over a long period of time. Expect SEO to take 6 months to 2 years.
TheRealPomax 26 mins If it takes you 6 months to 2 years to become prominent in local searches, something has gone very wrong. The answer to the original question is really one of "over how big a part of the internet?" first. Global SEO is a slow process, local SEO should be a matter of months at most, if the site you're optimizing for isn't "yet another one" in a saturated local market.
onion2k 18 mins How does that work when there are other companies also working on their branding? Why would one company's effort to create a good online brand work better than another company's effort? That is SEO - working out what has a material impact on search rankings and implementing it.
dx034 10 mins Is that because you need iterations to figure out what Google values or do they just punish new domains?
「首次公開代幣發售」（ICO，Initial Coin Offering）面世，並在多個國家或地區湧現，已有幾年歷史，尤其去年ICO在美國更刮起一股熱潮。ICO興起，各方反應不一，包括大小投資者以及監管機構等。有些國家禁止進行ICO，如中國和南韓。據中國人民銀行統計，內地發行的ICO，90%帶有欺詐性，故勒令立即停止。
究竟ICO未來發展命運如何？散戶值不值得投資ICO發行的代幣？ICO何以變成欺騙性交易？只有深入檢視ICO的交易本質，才有客觀判斷基礎。
ICO是一種融資活動。ICO的發起人，首先有一個生意「概念」，或一個發展項目構想，於是向外進行融資。傳統融資方法不外乎向銀行融資，或進行股票「首次公開發售」（IPO，Initial Public Offering），出售股權。不過，由於進行ICO所寄託的生意「概念」或項目發展構想，大部分只停留在「紙上談兵」階段，根本很難成功向銀行融資；也不能進行IPO，因為各國監管機構對於IPO都有嚴謹的上市規定。
項目概念紙上談兵
於是，ICO的發起人可另闢融資蹊徑，採用出售代幣（token）形式，以區塊鏈（blockchain）作為平台進行融資，一般在以太坊（Ethereum）進行，亦可在自行設置的區塊鏈上發行虛擬代幣。所發行的代幣與IPO買出的股權，兩者最大分別在於──小投資者在IPO購入的股份，是明確擁有公司的股權，為公司的小股東。但在ICO所購入的代幣，只能享用該公司未來所提供的服務。
很明顯，企業進行IPO，需要委託投資銀行準備上市文件，且公司要有實質性業務，也需要有盈利紀錄，需要符合上市規定，上市過程受到慎密監管，斷不能像ICO那樣，很多都沒有公司在實質性地運行、生意概念屬「紙上談兵」，遑論有實質性產品、服務和盈利；而且進行ICO也毋須遵照「訊息披露」守則，包括把公司的資料，如註冊地、股東身份、股權分布等向投資者披露；也不用根據上市守則，定期發表財務報表。唯一需要的，是準備一份進行ICO向外公布的「白皮書」。
這份「白皮書」需要披露什麼資料也沒有規定，包括公開最基本的公司營運資料、財務狀況、公司大股東或業務顧問是誰，所籌集代幣的具體用途等，只需由發起人自行「天才創作」。
另一方面，從IPO買入的股票，投資者是公司的股東，持有該公司的股權，擁有投票權，也擁有分享公司未來盈利的權利，但從ICO買入的代幣，投資者不僅沒有投票權，其所持並非公司的股份，只是購入了一個透過代幣去使用該公司未來提供的服務，不能分享公司日後成功發展而獲得盈利的一分一毫。可以說，購入的虛擬代幣，當該公司未來成功發展，獲得盈利，基本上跟你扯不上關係，只有該公司承諾未來所提供的服務增值，代幣的使用價值上升，投資者才能夠賣出代幣獲利，回報過程相當迂迴。
可以說，ICO的弊端相當顯著。第一，由於不受監管，使ICO出現「良莠不齊」現象，好壞項目難分，輕易成為欺騙性融資行為的溫床。
第二，雖然ICO投資資金的流動性高，作為投資者和公司始創人，可以在很早期便取得資金回報，但高流動性其實是一面「雙刃劍」。當始創人很快可以把初始的權益資本套現，就難免對公司缺乏長遠的發展籌算。反之，當所投資的資本需要待公司成功發展，取得盈利後上市才能套現賺錢，這樣，始創人便會更着眼於公司的長期發展。
多項弊端充滿投機
第三，投資資金高流動性，自然吸引到不少投機者進場，他們對公司的經營或發展項目是否成功並不關心，只着眼於「炒高」公司的股價或虛擬貨幣價格獲利，令ICO市場充滿投機之風。
第四，由於ICO不受監管，購入代幣的投資者為匿名，且持有量也不公開，基於交易量不算大，很容易受到「大戶」進行市場操控。此外，ICO虛擬代幣最大的持有量股東是誰？其持有量多少？何時有出售行為？完全缺乏透明度。由於區塊鏈平台買賣兩方都匿名，追查也查不出來。基於質素好壞難分，顯然一個典型的「檸檬市場」。
「檸檬市場」是2001年諾貝爾獎得主阿克洛夫（George A. Akerlof）教授在其開創性論文《檸檬市場：質量不確定性與市場機制》（The Market For “Lemons”︰Quality Uncertainty And The Market Mechanism）一文，分析在訊息不對稱（information asymmetry）下出現的「逆向選擇」（adverse selection）交易行為。
他以美國二手車市場為例。在美國，低質量車被稱作「檸檬」。在二手車市場，賣方對舊車質量所掌握的訊息比買方多出很多，於是出現訊息不對稱情況。但是，買方根據經驗，可以大致了解舊車市場商品的平均質量，他們願意以這個平均質量的預期價格水平購車。結果，高於平均質量的汽車，便被迫退出市場。由於高質量舊車退出市場，導致二手車市場上商品質量進一步下降，這樣一來，買方願意付出的購車價格水平，也隨之下降到新平均質量的水平，直至到只留下最低質量的舊車在二手市場交易，最後，由於市場規模逐步萎縮，整個市場「崩塌」。
可以看到，檸檬市場的交易是低效率的，改善這狀況方式之一，是有賴市場上的「誠信」。換言之，交易上的「不誠信」，需要付出社會代價。阿克洛夫指出，「欺騙性交易將誠實交易者逐出市場。市場上原本可能有買家想購買高質量商品，且有賣主願意在一個價格範圍內出售該種商品。
由此觀之，檸檬市場形成的條件包括：一、訊息不對稱，賣方所掌握的商品訊息比買家多；二、賣方有進行欺騙的積極性，以取得更大收益；三、商品質量的訊息透露，沒有一個權威性的中介作出確證，以致賣家可以把商品的質量說得「天花亂墜」。而上市公司的財務報表，必須得到核數師的核實，就是防止這種情況出現；四、商品或項目有好有壞，好壞往往難以區分；五、賣方把低質量商品以高質量商品的價格出售，沒有受到懲罰後果。
ICO市場完全符合這五項條件，清楚顯示出，ICO是一個檸檬市場的新典型，小投資者的確需要了解當中的交易風險！囿於篇幅，有關ICO好的一面，以及基於不誠實交易不僅將誠實交易驅逐出市場，還令市場規模不斷萎縮，因而監管實在有其迫切性，這些問題，筆者將另文討論。
本文由科大商學院傳訊部筆錄，黃昊教授口述及整理定稿
作者為香港科大商學院會計學系副教授
My Favorite PostgreSQL Queries and Why They Matter 217 grzm 8 hrs 46
https://severalnines.com/blog/my-favorite-postgresql-queries-and-why-they-matter
Unable to load the content https://severalnines.com/blog/my-favorite-postgresql-queries-and-why-they-matter
Facebook（fb）教主朱克伯格（Mark Zuckerberg）接受美國國會「公審」成為全球焦點，回看整個洩密風暴過程，可說示範了一間公司生死榮辱完全繫於一人身上的危與機，猶如當年阿根廷足球隊成也馬勒當拿，敗也馬勒當拿。在同股不同權企業快將破天荒在本港上市之際，事件值得香港股民借鑑。
這次fb東窗事發於上月19日，傳媒揭露這個全球最大社交網絡平台縱容第三方英國劍橋分析（Cambridge Analytica），挪用美國選民在fb填寫的個人資料，繼而左右2017年美國總統大選結果。根據fb後來自己披露，受影響用戶人數高達8700萬，佔美國總人口3.26億的兩成七，涉及人數之廣、影響之巨，堪稱史無前例。
預演不同權風險
遇上如此重大危機，作為上市公司掌舵人，按常理應第一時間站出來解畫甚至道歉，最低限度亦應發聲明交代補救措施，挽回公眾及投資者信心，然而朱克伯格卻選擇閉關四日，直到3月22日才現身解畫，但已導致fb股價連日急瀉，從3月18日收市高位185.09美元跌至3月26日低位149.02美元計算，最多插水兩成。
事隔四天才就重大事件回應，明顯不是危機管理的良好方法，而且Facebook董事局成員包括Netflix行政總裁黑斯廷斯（Reed Hastings），美國運通前行政總裁Kenneth Chenault和摩根士丹利前首席董事Erskine Bowles等精英，不可能沒有人勸他盡速向公眾交代，惟朱克伯格能夠依然故我閉關數天，大概全仗同股不同權制度。
據經濟師分析，朱克伯格經過連番沽售，現僅持有16%的fb股份，由於當中包含1股10票的B股，令這名33歲的創辦人擁有公司60%投票權，因此就算董事局內不乏fb主要股東，理論上可向主席勸諫游說，甚至「逼宮」，惟因朱仔享有「以小控大」優勢，即使主要股東連同散戶行動，以持有84%權益向朱仔施壓，但到最後投票階段頂多亦只有40%投票權；換言之，董事局「革命」尚未成功，將先被朱克伯格解散。
以小控大難逼宮
近日確有fb主要股東、紐約市退休基金投資的監督人紐約市審計官斯特林格（Scott Stringer）要求朱克伯格下台，但教主斷然拒絕，堅持自己是最適合的主席人選。
事實上，觀乎朱克伯格在4月11日及12日一連兩天出席美國聽證會的出色表現，其地位暫時難以被取代。尤其他在首天聽證會淡定「舌戰群儒」，獲投資者投下信心一票，fb當天收市飆4.5%，創2016年4月28日以來最大單日升幅。
目前在美國交易的科技股如Google母公司Alphabet、阿里巴巴和Snapchat母公司Snap等，皆採用同股不同權制度，其中以Snap最為離譜，公開發行的A股竟沒有投票權，而B股則1股1票；C股每股10票，由兩名創辦人Evan Spiegel和Robert Murphy持有，使兩人合計投票權高達88.6%。據聞二人IT味十足，不常與分析員或投資者溝通，令市場對Snapchat優化廣告配對功能的信心打上折扣，股價上市至今長期潛水，可見同股不同權是否成功，就要視乎「以小控大」的舵手之魅力、魄力和能力。
帶眼識人更重要
曾幾何時，阿根廷是馬勒當拿的「一人球隊」，1986年「上帝之手」淘汰英格蘭，1990年領軍勇奪世界盃，1994年因服禁藥被褫奪參賽資格、導致球隊16強出局。成也球王，敗也球王，有辣有唔辣，準備參與同股不同權遊戲的香港股民，細心揀股之餘，更重要的是帶眼識人！
hclee@hkej.com
美國三大銀行股周五（13日）公布業績，首季盈利均勝預期，美股早市先升後回軟。老特如果唔搞咁多嘢，讓股民回歸基本因素，則道指有機會借業績期扭轉劣勢【圖1】，但若然業績表現理想都無法扭轉大局，則麻煩矣。
美國三大銀行股摩根大通、花旗，以及富國周五晚上公布的首季度業績均勝過市場預期。今年首季金融市場十分波動，散戶嚇到半死，但銀行炒房見波動則風生水起，摩通的交易業務進賬唔少，其中股票交易收入上升了26%，債券交易收入亦增加8%，全季集團的整體盈利有87億美元，按年多35%。炒房業績有好轉，資金好快就湧去炒起摩根士丹利及高盛，因為這兩間華爾街大行一向炒賣業務都好大。
市場起伏大 炒房風生水起
據摩通對2018年的盈利計算，已高於2017年高峰三分之一，其中包括了減稅的得益，行政總裁戴蒙（Jamie Dimon）話今年會把部分盈利投資於科技，興建新總部及增設400間分行。銀行業乃百業之母，銀行資產負債表強勁，放貸亦會鬆手點，這將有利整體的經濟增長動力。
微觀企業盈利增長可觀，但老特的政策仍會引發不少動盪，投資也最好取一個平衡，不應全情投入。美國打唔打敍利亞？中美貿易戰如何演變？兩大疑慮還是懸而未決。有人會話這些都只是噪音，還是聚焦於經濟前景及企業盈利更實際，而且特朗普五時花六時變，很難跟着他的推文做投資決定。不過，狼來了的故事教訓大家狼始終會來，狼一來就會雞飛狗走。
老特敍利亞或搞大龍鳳
老特話打敍利亞，不過遲遲未動手，有報道引述白宮幕僚說原來老特是想搞一場大龍鳳，不單要教訓敍利亞巴沙爾政府，還要打擊他背後的支持者伊朗及俄羅斯在當地的軍隊，他嫌國防部擬備的方案殺傷力不夠，國防部則恐怕如果傷及俄國人員，可能爆發美俄軍事大衝突，老特似乎比軍人更想大打一場，敍利亞風波最後點收科仍是一個未知數。
至於中美貿易戰，就更加唔好當無到。據華爾街日報引述華府官員表示，貿易代表處最快會於下周公布千億美元的制裁中國貨品清單，由於金額龐大，今次的清單將很大機會包括日常消費品如成衣、波鞋及手機等。美國財政部亦密鑼緊鼓起草規例嚴限中國企業透過收購、合資、特許經營或是其他安排投資於美國的先進科技。
從種種跡象中顯示，中美的貿易戰仍未真正開始，內地也要顧慮，中方愈退，美方會愈得寸進尺。美國對新興勢力中國的猜疑愈來愈大，惟恐中國會在世界經濟中超越美國，必須先下手為強，箝制中國。
中美貿易戰鼓聲會再響
老特周五忽然轉軚，提到會研究重新加入跨太平洋夥伴關係協定（TPP），條件之一是其他成員國須給予較奧巴馬年代為佳的待遇，很明顯，這不易談，但老特欲透過TPP，聯合其他國家在經濟上圍堵中國，用意都幾明顯，所以，若說中美的經貿關係已因習大大在亞洲博鰲論壇的改革開放演詞就得到解決，未免過於樂觀。
老特對中國開放市場及減關稅的時間表遠較中國進取，進口車的關稅最好是一減到底，外資進入中國多個行業最好是立刻進行，而非中方如意算盤的三年分期執行，很多時事情到埋枱談判時，矛盾便會浮現。
中國擴大對外開放，對本地產業會有壓力，外來競爭增多，企業的利潤率自會受損，近期內地滬深300指數表現疲弱，20天移動平均線跌得仲急過恒指【圖2】，可能正正反映緊這方面的憂慮。
4月12日，周四。Facebook教主朱克伯格（Mark Zuckerberg）先後出席美國國會參眾兩院聽證會，首輪接受參議員質詢，全世界當笑片嚟睇，難為朱仔要扮嚴肅，以免態度輕浮罪加一等。裝作老成持重唔難，忍笑先戥佢辛苦。
老畢有睇首輪答問直播，次輪卻只看媒體報道，朱克伯格在眾議院算唔算過關，感覺不像第一輪那麼「真實」。無論如何，朱仔已闖過國會山莊木人巷，重回Facebook大本營，不妨跟大家分享一下個人對這場大龍鳳的看法。
代表股東提問？
不必報上年齡，睇樣已知向朱克伯格提問的參議員，好多都做得佢爺爺嫲嫲。84歲的猶他州參議員Orrin Hatch質問朱仔，Facebook不向用戶收費，如何經營下去？網上所以瘋傳聽證會這一幕，只因問題本身已暴露了Hatch對社交媒體一無所知，搞到朱仔呆了數秒才「回魂」，冷面笑匠般應之曰：「Senator, we run ads.」朱仔唔可以笑，但網上網下已當正佢係星爺。
老畢當然也有份笑，可是轉念一想，84歲老人家不懂社交媒體營運模式，何奇之有？與其笑這位元老級政客「科技盲」，不如問一問：若非事先知道這是參議院聽證會，你會否以為周二開的乃Facebook AGM，席間一位年老股東質問做得他孫子的CEO，公司不收費如何能賺錢？姑勿論Hatch的問題白癡不白癡，從出發點着眼，提問者關心的是Facebook不收費憑什麼經營下去，怎看也是站在股東利益要求朱克伯格解畫。
聽證會有預先設定的agenda，幾十名參議員聚首一堂所為何事，跟年紀、懂不懂科技無關，何以Hatch好像關心Facebook不收費便難以替股東創造價值，多於大會白紙黑字寫明的濫用數據侵犯私隱？
爺爺錫住小朱
另一個有趣的地方是，參院聽證會以司法及商務委員會聯席名義召開，意味具深厚法律背景的議員約佔半數。律師不可能凡事皆懂，但處理的案件卻經常涉及不同範疇專業，除了搜證不容馬虎外，對己避重就輕、對敵咄咄逼人，是一種把勝算推向自己的重要技巧，而不斷要求對方正面回答Yes或No，往往有助達到這種效果。
從當天直播發現，不懂科技但懂法律的參議員不少，惟由始至終不見有人以嚴峻語氣逼朱克伯格清楚回答Yes或No。法律界出身的議員如此「錫住」小朱， Facebook股價怎會不一路問一路升？
眾議院在下並未現場觀戰，但從各方報道得知，有議員用此招試圖逼小朱埋牆角，他想帶發問者遊花園，對方便打斷其話頭，要求朱仔直接回答Yes或No。
其中一位議員問朱克伯格：「你是否願意為了保護用戶私隱而改變Facebook的商業模式？」老畢心想，即使是一個普通記者，只要稍為了解社交媒體，都懂得問這條「尖銳」問題。不難想像，以朱仔上國會山莊前演練之充足，此類涉及Facbook生意與用戶私隱之間重大矛盾的問題，要應付該不難。然而，朱克伯格非但沒有直接答Yes或No，且以「不明白閣下問什麼」（I'm not sure what that means）作答。小朱不敢觸碰這個話題，足證答Yes又死答No又死，惟有扮唔明博大霧。
用戶現在要回頭已很難，馬斯克（Elon Musk）大大聲話刪除Tesla的fb專頁，蘋果老總庫克（Tim Cook）又同朱仔隔空對罵，但科技公司其實坐埋同一條船，加強監管唔會只針對Facebook。
Tesla傳破產，Model 3又趕唔上進度，馬斯克要頭痕的問題多籮籮，居然仲有時間身體力行delete Facebook，得閒得滯乎？「鐵甲奇俠」點會死錯人，趁朱仔畀人插緊轉移視線罷了。
科網巨頭的問題不在用戶怕私隱外洩，集體改變已經改變唔到嘅生活習慣，而係眾巨頭已成眾矢之的，自我監管行不通，就由外部力量加以限制，一旦觸及數據的蒐集、使用、分享和儲存，FANG經營成本會增加，業務增長會受壓；早已當正四家公司可以控制全世界的投資者，還會不會給予它們「世界級」的估值？
房屋委員會「出售居者有其屋計劃單位2018」(下稱新居屋)3個全新屋苑共4431個單位，昨天截止申請，房委會數據顯示，截至昨天下午7時累計收到約16.6萬份申請，較推出單位超額逾36倍。
同日截止接受申請的白居二2500個名額，則累計收到約3.9萬份申請，超額約14.6倍。
發言人表示，由於大部分綠表申請須經屋邨辦事處核實資格，然後才送交居屋銷售小組登記；加上在截止前經郵遞送交的申請表仍未完全送抵，故此上述數字仍未反映最終的申請數目。
新居屋預計將於6月進行攪珠，屆時公布最終的申請數字。
今期新居屋分布於九龍東啟德啟朗苑、長沙灣凱樂苑和東涌裕泰苑，實用面積278至631方呎，以市價七折定價，售價159萬至630萬元。
Many developers view monetization and design as two separate entities. But based on our experience at Tappx, they should have monetization deeply rooted at the starting point when building an app.
But there are many factors to consider when picking a strategy. So, we’ve listed down a few strategies and their unique qualities to help you evaluate which will be the most compatible with your app/business.
In-app purchase (IAP) For the in-app purchase model, you need to find the proper balance between what content or functionality will be included in the base download vs what can be unlocked with payment. If too little content is provided, users will look elsewhere for similar products. On the flip side, users won’t have a reason to upgrade if you give too much free content.
For game developers, consider offering both fixed-price and in-game currency options for unlocking content. This allows non-paying users to experience unlocking “unlockable” content while allowing paying users to freely spend via both methods. Alternatively, you can employ video ads to reward users with “free” currency, thus obtaining ad partners and user allegiance.
How to implement it No matter what your app is, you’ll need to deploy and scale a large variety of premium options to offer your audiences at different price tiers. You can do this by defining a pricing strategy.
Be sure to communicate the value of your premium offerings at the right time by discovering the maximum engagement moments for your users and then impacting them with paid offers. Take advantage of urgency and offer discounts as well. You can communicate to users via in-app messages or through push notifications.
You can also offer users a “taster” experience of what it feels like to be a user with premium capabilities. If you use ads, you may use them to offer paid offerings to your users.
In-app advertising In-app advertising consists of distinct categories: video ads, display ads, and native ads. By implementing ads, you can focus on coding and improving your game. On the other hand, this allows users to download apps for free without restrictions on content or functionality other than the ads themselves.
When choosing how to monetize your app via advertising, you should choose an ad partner that will help you find the right advertising strategy and ad format for your app.
How to implement it Putting ads in your app is easy, but getting good revenue from them is challenging.
You can start by selecting an ad partner. Then, select the ad format you’ll implement. Depending on these, your revenues can fluctuate a lot. Try to find a balance between the frequency your ads will show and the session length. Keep the user at the center of the experience.
Take into account that eCPM (effective cost per thousand impressions) is not the main metric to keep an eye on. This is because the metric depends on other factors, some of which you cannot control. Keep monitoring ARPDAU (average revenue per daily active user) data, check your metrics regularly, and segment your user base. Based on these, set up an ad strategy accordingly, and don’t forget to do extensive AB testing.
Subscription-based Although subscriptions only accounted for 4 percent of mobile app revenue in 2017, they can be highly lucrative for the right kind of app (e.g. Netflix, Amazon, Spotify). Another type of subscription lets users access the full functionality of an app with a recurring charge.
Giving users a free trial to your subscription-based app is a common method to allow them to fully experience the benefits of your app before committing to the service. Subscriptions often have cheaper monthly pricing schemes than fixed-cost apps, thus making for a more affordable alternative.
Subscription success relies on the quality of the service, so it can be a good way to gauge your audience’s response to features or content. Developers should try to release a steady stream of new content, along with regular app updates to keep audiences engaged and subscriptions renewed.
(Note: Paid users will be more discerning than those who download free apps on a whim and will demand greater levels of customer service and support.)
How to implement it If you’re looking to integrate this model, it’s important to know what the premium features are going to be and the reasons why users need to pay for them. It’s a risk to offer features for free. So, one of the first things you need to do before implementing a subscription model is to ascertain what is valuable for your users when they use your app and then execute a trial plan for them to experience what it’s like to be a premium user.
The objective is to make users feel that everything will be the same for them in your app after the trial. Also, take into consideration that users are different and have varying needs. So, it’s recommended that you develop a subscription plan based on different price tiers and features, according to your different segmented audiences.
Direct price The direct price model (aka pay-per-install model) can be divided into two categories: paid and “paidmium” apps. The former refers to apps that have a one-time download cost while the latter refers to paid apps that also offer in-app purchases. For this model to work successfully, your app needs to have original content that cannot be found in free or copycat apps.
Due to the heavy saturation of the app market, this model has dropped in revenue and popularity. Competition with apps that allow users to test products before investing in it has caused fewer developers to go for this model.
How to implement it Ensure first that your value proposition is unique and that it deserves to be premium. Study what the competition is doing and what prices they are offering. Then, create your pricing structure. Take into account the elasticity/seasonality in the year to maximize revenues and downloads.
Don’t think of pricing as fixed at one level; it isn’t. It can be changed according to the evolution of your products or according to the feedback of your users. Understand the market demands and never forget that the price should be sustainable for your business in the long term.
Decide which monetization model to use from the very beginning. Then, test and improve along the way. Study your users, observe how they interact with your product, and try new improvements. Ultimately, the user relationship with your app will dictate your product’s success.
Stay a step ahead with the upcoming Tech in Asia Premium.
Committed to producing deeply reported content on tech and startups in Asia, we will be moving into a paid subscription model in the near future. This will allow us to continue investing in creating content of value for you.
Secure an early bird rate today when you pre‑register.
LEARN MORE Editing by Jaclyn Teng
(And yes, we’re serious about ethics and transparency. More information here.)
Community Writer Kinga Kruzmanowska Publisher Development Director at Tappx. VIEW ALL COMMENTS
港產配對補習導師服務平台Snapask宣布，完成1500萬美元（約1.17億港元）的A輪融資，由泰國正大集團家族成員謝展領投是次融資。
據Snapask表示，是次融資所得資金將用作全力發展機器學習（Machine Learning）和招募數據科學人才，公司現正着手開發全新「學習助理」（Learning Planner），計劃透過統計模型分析出學生的弱點與長處，再利用機器學習做到資料分群，並基於學生所提供的學習歷程建立學習模型。
今次並非Snapask首次成功融資，去年5月公司宣布完成Pre-A輪融資，獲注入500萬美元（約3900萬港元），該次融資已獲風險投資公司Kejora Ventures、微光創投（由騰訊集團前首席技術官張志東及前高級執行副總裁吳宵光創辦），以及美圖（01357）董事長蔡文勝等注資。
Snapask成立初期，主要透過手機應用程式Snapask提供一對一功課問答服務，發展至現時已與50間香港及台灣中學合作推數據為本學習，透過與學校合作收集前線教學經驗及意見，並與日校老師分享學生課後學習數據，讓老師可以更全面掌握學生課後學習情況。
騰訊領投社交電商拼多多
另一方面，內地媒體報道，內地社交電商「拼多多」近日完成新一輪30億美元（約234億港元）的融資，估值接近150億美元（約1170億港元），由騰訊（00700）再次擔當領投角色，紅杉亦參投。早在2016年7月，騰訊已在「拼多多」的1.1億美元B輪融資參與投資。
據了解，成立兩年多的「拼多多」，去年商品交易總額（GMV）已超過1000億元人民幣。有分析指出，騰訊再度加注，或反映零售領域對其戰略意義愈來愈重要。
48 ussumant 4 hrs 32
Looking for something where the Value of the Product is shown in real time like Grammarly's Text correction. fulafel 2 hrs The latest US presidential election.
pesenti 43 mins http://densepose.org/ Which can extract the surface of the human body in real time.
ericjang 3 hrs Tacotron 2 text-to-audio samples:
https://google.github.io/tacotron/publications/tacotron2/ind...
Blog post: https://research.googleblog.com/2017/12/tacotron-2-generatin... Paper: https://arxiv.org/abs/1712.05884
philprx 2 hrs GitHub GIT or it doesn't exist ;)
jcmeyrignac 2 hrs https://www.deepl.com/translator
alex_duf 25 mins I just tried the french to english. It's better than google translate but not that good.
arunbahl 3 hrs WordLens blew me away when it launched in 2010 (now part of Google Translate). https://en.wikipedia.org/wiki/Word_Lens
zawerf 3 mins I thought augmented reality with mobile camera was going to take off after seeing that demo. But 8 years later and all we have are snapchat filters (and pokemon go for a brief few weeks).
Hopefully ARKit/ARCore will make it easier to create the next the-future-is-now level of awe-inspiration app.
WorkLifeBalance 29 mins It should have been a killer-app for google glass.
Had google glass not stored video and processed everything in real-time it could have avoided the "glasshole" stigma.
People would have likely mocked it for not being able to capture video, something they could have added in later in "response to demand".
The world doesn't need more omni-present recording devices, but solid AR would be a net benefit.
akerro 27 mins DuckDuckGo and Google Search probably.
MiniCreo 3 hrs This site uses AI to enhance your low-res photos: https://letsenhance.io/
rtcoms 2 hrs Google photos
Face and object recognition helped me quite a lot in finding old photos.
vbsteven 2 hrs Spotify daily mixes
tontonius 1 hr would you care to elaborate on what parts of Discover that are powered by AI?
In all honesty, I think the lines are severely blurred at this point.
blixt 1 hr AI has been a moving goalpost since its conception, all the way from
• "it can play tic tac toe";
• to "it can beat people at chess, Jeopardy, Go, DotA";
• to "it can see, understand and physically move around our world";
• to "it solves problems the smartest humans can't begin to fathom"
Spotify's machine learning algorithms are definitely beyond tic tac toe as they can learn from millions of listeners and look at a single listening history to figure out great music suggestions. I would put that under the AI umbrella as it's something people were doing before computers did it.
GrumpyNl 1 hr Is that AI from spotify or just smartly compairing lists?
blixt 1 hr What is "smartly comparing lists"? To describe further (as I'm an ex-Spotify employee I have some outdated insight into how it works):
Spotify has historically used machine learning to tweak a predictive engine that can convert a track or artist into an N-dimensional value and then use the distance to other tracks/artists in this N-dimensional space.
Is that AI? Maybe. How does the brain work? Maybe when you see a dog it's converted into an N-dimensional space where cats are pretty close, at least much closer than turtles. So if that is human intelligence, is Spotify's recommendation engine not artificial intelligence?
denzil_correa 1 hr Microsoft's Skype Real Time Language Translation (2014)
https://www.youtube.com/watch?v=RuAp92wW9bg
purplezooey 1 hr Eliza
louismerlin 3 hrs https://experiments.withgoogle.com/ai/teachable-machine
The teachable machine from Google is a great little experiment.
neel8986 3 hrs Google search, translation?
cm2012 2 hrs Facebook's newsfeed is basically a superimpressive AI that learns from what people click and maximizes ad revenue from that.
Spearchucker 1 hr That same AI drove me away. I subscribe to 3 hobby-related groups and because virtually all of my family and friends no longer post anything on Facebook, I only ever see content related to that hobby.
majewsky 1 hr If that description is honest, why do you blame the algorithm for selecting hobby-related content from a set of inputs that only contains hobby-related content? That's like buying a football magazine, then complaining that it's only football, football, football in there.
kotapi 3 hrs Cozmo https://www.youtube.com/watch?v=DHY5kpGTsDE
oceanman888 1 hr I personally think the impressive part of cozmo comes from interaction and design, rather than AI.
robax 3 hrs Ambient.ai has a very demonstrative landing page. https://ambient.ai
jawrainey 2 hrs There's an object detection library (YOLO1) freely available that achieves similar bounding boxes/matches to ambient's homepage demo; Joseph Redmon also gave a TedX2 about it that describes some of the technical details.
nstj 1 hr I just saw the YOLO stuff in action at AWS Summit and it was epic. Super fast and really really awesome.
taneq 2 hrs Is there meant to be something on the second 'page' (ugh I hate this single-page-permascroll-with-fixed-header layout that they all do these days) other than the 'let there be more light' box? It's just a grey background so I suspect something didn't load there. Otherwise I just see a generic video showing object tagging on the first 'page' and the pretty-but-pointless graph thing on the third 'page'.
Firefox 59.0.2 with PrivacyBadger and uBlock Origin (although I turned both off and still nothing loaded).
jackbrown77 3 hrs Both siri and google assistant are very good AI's, but there is lot to explore in the industry of AI.
arvigeus 2 hrs Not Hotdog app
MiniCreo 3 hrs Silly Siri
美股周二（10日）續向上，三大指數均漲逾1%。儘管如此，屬長期「大淡友」的對沖基金經理赫斯曼（John Hussman）最近發表評論文章再度提出警告，認為股市近期出現如此劇烈震盪，進一步加深他對後市悲觀的看法。赫斯曼估計，標指將至少回落60%以上；而未來10年或更長時間，股市的回報將會是零，甚至是負數。
本欄上月底〈美股震幅超去年 重拾升軌無咁易〉一文曾指出，股市波幅擴大是不可忽視的因素。進入第二季度不久，波動情況依然頻密。截至上個交易日止，標普500指數已合共出現28次逾1%的變幅（15次升、13次跌），而對上一次如此波動的市況是2016年。回看去年，只有8個交易日出現逾1%的收市變幅；然而，今年僅第一季度已出現23次，接近去年全年的3倍。另外兩個主要指數的情況亦近似，道指及納指今年以來分別有29個和30個交易日出現逾1%變幅，是2009年以來最多。
更甚的是，標指去年未曾出現逾2%的變幅，今年卻已出現了8次，當中只有一次是上升，其餘皆是下挫，2月5日的跌幅更超過4%。而向來相對波動的納指，在去年只有3次2%以上的變幅（一升兩跌），但今年迄今已出現了10次（3升7跌，其中兩次下跌幅度逾3%）。
此外，美國芝加哥期權交易所波動指數（CBOE Volatility Index, VIX）今年也幾乎翻了一番；近期徘徊於20%這個長期平均水平附近，呈橫行待變格局。VIX理論上是反映期權交易商預期標指未來30天的波幅，同時也有「恐慌指數」之稱，今年初以來合共出現6次飆升逾20%，其中2月初更曾暴漲逾倍。在風聲鶴唳的市況下，難免令投資者異常不安。
赫斯曼認為，以潛在的貿易戰甚至美國聯儲局對利率前景取態轉變等新聞用作衡量市場風險，在某程度上是無關重要。
他指出，聯儲局的量化寬鬆政策（QE）是抬高股市估值到不可持續水平的最大力量；惟隨着接近零成本的廉價資金逐漸消失，一旦下跌漩渦形成，將會令更多投資者尾隨。赫斯曼認為，投資看股票估值，投機看市場心理，目前上述兩個因素都不利後市發展。
無疑，美股經歷2月至3月多次急挫，多少也會令投資者感到恐慌，但實際上仍有一些令人信服的理由堅持繼續持有股票。Facebook行政總裁朱克伯格上日在美國參議院作證，力求避免國會立法監管公司營運，股價隨之反彈；普遍科技股走勢似乎有穩定下來的跡象。再者，根據FactSet數據顯示，美股第一季度企業盈利增長應該高於17%，為2011年第一季度以來最高；而去年底預測僅為增長11%。這意味着儘管過去3個月的宏觀環境存有很大的不確定性，但企業盈利前景看起來仍然不俗。
港股方面，恒生指數收報30897點，升168點（0.55%），主板成交金額1329.77億元。從技術走勢來看【圖】，恒指重返由2017年向上伸延的上升通道底線之上；如能企穩，料有機會進一步回穩甚至造好。
