Following these proceedings, Macquarie upgraded its iQIYI Inc ADR [IQ] rating to a Neutral from a an Underperform in a research note published on Tuesday. PT values the company’s stock at a premium of 8.98 to its Tuesday closing price. On Tuesday ...

iQIYI Inc ADR (IQ) receives a Neutral rating from Macquarie

The stock of iQIYI Inc ADR (IQ) has gone up by 4.35% for the week, with a 3.45% rise in the past month and a -24.29% drop in the past quarter. The volatility ratio for the week is 5.60%, and the volatility levels for the past 30 days are 4.

iQIYI Inc ADR (IQ) Stock: Assessing the Risk and Reward

PubMatic , an independent technology company, has announced the availability of its newest offering, Activate, in the Asia-Pacific region. PubMatic’s new end-to-end supply path optimization (SPO) solution allows buyers to execute non-bidded direct deals on PubMatic’s programmatic platform,

PubMatic brings Activate to APAC

The service allows buyers to execute non-bidded direct deals on PubMatic’s programmatic platform, accessing premium video and CTV inventory at scale

PubMatic introduces SPO solution offering 'Activate' in APAC

These are five stocks for aggressive investors looking to get share count leverage on companies that have sizable upside potential. While not suited for all investors, they are not penny stocks with absolutely no track record or liquidity.

Snap and 4 Other Well-Known Stocks Under $10 With Huge Upside Potential

These are five $10 stock picks for aggressive investors looking to get share count leverage on companies that have sizable upside potential.

Global listings courtesy of LivesoccerTV: (*If you don’t see your country listed – You can email info@livesoccertv.com and ask them whether there is a channel showing Wolves v Newcastle live TV, or if they have their info wrong on any country/channel, please let them know via that email address)

Watch Wolves v Newcastle Live TV – These global channel listings for Saturday

To enhance the authenticity of the scenes, the production team finally selected Jijiazhai as the main filming location. Situated in Yongjia county of Wenzhou city in Zhejiang province, Jijiazhai is an ancient village dating back to the Song Dynasty and is honored as "China's landscape village".

Online TV drama portrays idyllic ancient rural life

In the current digital economy era, emerging technologies such as cloud computing, big data, the Internet of Things (IoT), and 5G are flourishing at an unprecedented pace. These advancements assist numerous enterprises in achieving cost reduction and efficiency enhancement through digital transformation.

China’s Zero Trust Network Access: Software-Defined Perimeter Market Worth Approximately $139 Million in 2022

Tech stocks led Asian equities traded in the US as American depositary receipts sharply higher Tuesday morning, rising 1.29% to 1,582.69 on the S&P Asia 50 ADR Index. From North Asia, the gainers were led by computer hardware company Canaan (CAN) and online game developer The9 (NCTY),

Tech Stocks Lead Asian Equities Traded in the US as American Depositary Receipts Sharply Higher in Tuesday Trading

iQIYI

iqiyi

Paid streaming content has overtaken the media industry space, with industry incumbents like Netflix and Hulu making a significant impact on the way content is consumed throughout the world, resulting in customers transitioning from traditional cable and satellite-based models to these paid streaming services. Originally offering consumers a way to reduce their expenses, the number of streaming platforms offering exclusive content has ballooned to the point where streaming services’ original value proposition has become eroded, with streaming bundles pricing equivalent to previous cable contracts. With this in mind, the next disruption to the media creation space will leverage Web 3.0 principles to reflect a new continuum: free streaming media services for all. now reflecting Why free media? One of Web 3.0’s primary characteristics is providing decentralized infrastructure with the goal of increasing peer-to-peer interactions, enabling more inclusion of traditionally marginalized populations in everyday digital life. Additionally, Web 3.0’s decentralized blockchain protocol will help individuals regain control over their time and data while being properly compensated, in contrast to today’s world where multinational enterprises are the ones profiting from individual consumer data. Increased participation in content consumption leads to potential benefits for both organizations and individuals alike; therefore, individuals having formalized equitable access to content makes sense. A similar parallel is observed in the wave of lawsuits targeting peer-to-peer music sharers in the , where companies eventually found their own ways to change their business models to adapt to consumer tastes. This resulted in the current framework of music streaming at no cost, with consumers choosing to pay for enhance features such as ad-free experiences or the ability to download music for offline use. early 2000s Free media is also important given the newfound societal emphasis of ensuring equitable access to resources for all. The ever-growing focus on equity and inclusion by increasingly younger demographics is particularly prevalent, with Gen Z specifically as the sustainability-minded generation. Free streaming media services provide a way for consumers to consume content on an equitable basis, ensuring the inclusion of anyone throughout the world as a member of the ever-growing global community. making their mark How would this work? The Harvard Business Review the goal of a business as “[to] make  a profit so that the business can do something more or better.” In this specific example, the boom of streaming media services has led to an increase in original content production, with consumers distinguishing between services based on factors such as content availability, originality, and prospects. In this case, a world with free streaming media services would be ad-enabled, where users watch a succession of short advertisements in exchange for formalized free content. describes Some examples of comparable successes within the space include Amazon’s IMDb TV and Baidu’s iQIYI, with the former being ad-supported and the latter offering a limited experience compared to its paid premium “VIP” service. Samsung's TV Plus, also provides a free alternative for consumers wishing to watch television with no cost, though this feature involves the purchase of an actual Samsung television to access Samsung's native TV Plus service. However, the potential of Web 3.0 is more than just consumers watching advertisements or receiving limited platform access. The world of non-fungible tokens (NFTs) can enable better participation within a free streaming media ecosystem. In this case, users who choose to participate and interact with advertisements can be rewarded for their actions, with the “watch-to-earn” model folding well into other . Through participation and interaction with advertisements, users can be rewarded on a normalized basis through digital rewards, which could then be exchanged for either tangible goods or to skip ads for a duration of time, among other things. “labor-to-earn” models One important point to note is that this hypothetical free steaming media  model would not involve services shuttering their paid services model. Instead, consumers can opt-in to pay streaming services for the benefit of  an ad-free experience, thereby maintaining this source of revenue for large organizations. Through utilizing “watch-to-earn” models to incentivize users to watch and participate within advertisements, companies can still maintain healthy profits while providing content free-of-charge to the rest of the world. Test cases of this phenomenon Free-to-use social media platforms like YouTube and Instagram provide two excellent test-cases of this phenomenon. These current examples can be seen through YouTuber Donovan Price and Instagram star Wayne Colley, with both these content creators providing specific use-cases of free content that is ad-supported, ensuring benefits for the creator, the YouTube platform, and advertisers alike. YouTuber Donovan Price, also known online as , has transformed the content creation ecosystem, with his series #BlackChicagoBeLike evolving over the past few years from 3-5 minute videos describing life in Chicago to 20-minute sagas around himself as the titular character. Korporate Bidness releases content akin to the traditional paid TV cycle, with his content releasing several times per month with a video length similar to television sitcoms. Korporate Bidness Similarly, Wayne Colley, also known as , provides another excellent example into the free content creation space. Based in Atlanta, Kountry Wayne creates storyline-quality content on three predominant characters, weaving hilarious and relatable content in the form of himself, a hustler named Drip, old-school player named Buddy, among other characters. However, Colley’s content is released on a nearly daily basis, with these three-to-five-minute segments totaling to nearly 30 minutes of weekly content creation on multiple simultaneous storylines. Kountry Wayne Both Price and Colley can generate revenue based on advertisements via the YouTube and Instagram platforms, monetizing their videos for profit. Additionally, both platforms benefit with an increased userbase, leading more consumers to spend time on that specific platform searching for similar content. Lastly, advertisers can still maintain profitability in this model, advertisers would potentially receive increasing amounts of revenue through the targeted nature of platform-based advertisements. This would be compounded with additional factors such as user search history, data attribution models, and other tools enabling increased engagement with ads. Conclusion While streaming media has transitioned consumers from a paid television model to paid web services, Web 3.0 will result in the formalization of free streaming for all. Pirate sites and other informal avenues already exist for content consumption, with these platforms typically only benefiting the illegal platforms themselves, to the great annoyance of a frugal consumer forced to click through ads and advertisers preying on clickability, as opposed to actual click-through purchases. A free steaming media services model enables a better ecosystem for consumers, platforms, and advertisers alike, providing a formalized way for consumers to control their time and data, while being rewarded for watching content they enjoy.

How the Future of Streaming Will Leverage Web 3.0 To Make Media Free of Every End User

If you couldn’t make it to , no worries. Below is a list of top 10 papers everyone was talking about, covering DeepFakes, Facial Recognition, Reconstruction, & more. CVPR 2019 --------------------------------------------------------- 1.Learning Individual Styles of Conversational Gesture Full Paper: https://www.catalyzex.com/paper/arxiv:1906.04160 . TLDR: Given audio speech input, they generate plausible gestures to go along with the sound and synthesize a corresponding video of the speaker : Speech to gesture translation model. A convolutional audio encoder downsamples the 2D spectrogram and transforms it to a 1D signal. The translation model, G, then predicts a corresponding temporal stack of 2D poses. L1 regression to the ground truth poses provides a training signal, while an adversarial discriminator, D, ensures that the predicted motion is both temporally coherent and in the style of the speaker Model/Architecture used : Researchers qualitatively compare speech to gesture translation results to the baselines and the ground truth gesture sequences (tables presented by the author show lower loss and higher PCK of the new model) Model accuracy Speaker-specific gesture dataset taken by querying youtube. In total, there are 144 hours of video. They split the data into 80% train, 10% validation, and 10% test sets, such that each source video only appears in one set. Datasets used: --------------------------------------------------------- 2.Textured Neural Avatars Full Paper: https://www.catalyzex.com/paper/arxiv:1905.08776 TLDR: The researchers present a system for learning full-body neural avatars, i.e. deep networks that produce full-body renderings of a person for varying body pose and camera position. A neural free-viewpoint rendering of human avatars without reconstructing geometry : The overview of the textured neural avatar system. The input pose is defined as a stack of ”bone” rasterizations (one bone per channel). The input is processed by the fully-convolutional network (generator) to produce the body part assignment map stack and the body part coordinate map stack. These stacks are then used to sample the body texture maps at the locations prescribed by the part coordinate stack with the weights prescribed by the part assignment stack to produce the RGB image. In addition, the last body assignment stack map corresponds to the background probability. During learning, the mask and the RGB image are compared with ground-truth and the resulting losses are backpropagated through the sampling operation into the fully-convolutional network and onto the texture, resulting in their updates. Model/Architecture used : Outperforms the other two in terms of structured self-similarity (SSIM) and underperforms V2V in terms of Frechet Inception Distance (FID) Model accuracy : Datasets used - 2 subsets from the CMU Panoptic dataset collection. - captured our own multi-view sequences of three subjects using a rig of seven cameras, spanning approximately 30 degrees.                                          - 2 short monocular sequences from another paper and a Youtube video --------------------------------------------------------- 3.DSFD: Dual Shot Face Detector Full Paper: https://www.catalyzex.com/paper/arxiv:1810.10220 TLDR: They propose a novel face detection network with three novel contributions that address three key aspects of face detection, including better feature learning, progressive loss design and anchor assign based data augmentation, respectively. : DSFD framework uses a Feature Enhance Module on top of a feedforward VGG/ResNet architecture to generate the enhanced features from the original features (a), along with two loss layers named first shot PAL for the original features and second shot PAL for the enchanted features. Model/Architecture used : Extensive experiments on popular benchmarks: WIDER FACE and FDDB demonstrate the superiority of DSFD (Dual Shot face Detector) over the state-of-the-art face detectors (e.g., PyramidBox and SRN) Model accuracy :  WIDER FACE and FDDB Datasets used --------------------------------------------------------- 4.GANFIT: Generative Adversarial Network Fitting for High Fidelity 3D Face Reconstruction Full Paper: https://www.catalyzex.com/paper/arxiv:1902.05978 TLDR: The proposed deep fitting approach can reconstruct high-quality texture and geometry from a single image with precise identity recovery. The reconstructions in the figure and the rest of the paper are represented by a vector of size 700 floating points and rendered without any special effects.(the depicted texture is reconstructed by the model and none of the features taken directly from the image) : A 3D face reconstruction is rendered by a differentiable renderer. Cost functions are mainly formulated by means of identity features on a pretrained face recognition network and they are optimized by flowing the error all the way back to the latent parameters with gradient descent optimization. End-to-end differentiable architecture enables us to use computationally cheap and reliable first order derivatives for optimization thus making it possible to employ deep networks as a generator (i.e,. statistical model) or as a cost function. Model/Architecture used : Accuracy results for the meshes on the MICC Dataset using point-to-plane distance. The table reports the mean error (Mean), the standard deviation (Std.).are lowest for the proposed model. Model accuracy :  MoFA-Test, MICC, Labelled Faces in the Wild (LFW) dataset, BAM dataset Datasets used --------------------------------------------------------- 5.DeepFashion2: A Versatile Benchmark for Detection, Pose Estimation, Segmentation and Re-Identification of Clothing Images Full Paper: https://www.catalyzex.com/paper/arxiv:1901.07973 TLDR:  Deepfashion 2 provides a new benchmark for Detection, Pose Estimation, Segmentation and Re-Identification of Clothing Images : Match R-CNN contains three main components including a feature extraction network (FN), a perception network (PN), and a match network (MN). Model/Architecture used : Match R-CNN achieves a top-20 accuracy of less than 0.7 with ground-truth bounding boxes provided, indicating that the retrieval benchmark is challenging. Model accuracy : The DeepFashion2 dataset contains 491K diverse images of 13 popular clothing categories from both commercial shopping stores and consumers Datasets used --------------------------------------------------------- 6.Inverse Cooking: Recipe Generation from Food Images Full Paper: https://www.catalyzex.com/paper/arxiv:1812.06164 ages TLDR: Facebook researchers use AI to generate recipes from food im : Recipe generation model- They extract image features with the image encoder. Ingredients are predicted by Ingredient Decoder, and encoded into ingredient embeddings with ingredient encoder. The cooking instruction decoder generates a recipe title and a sequence of cooking steps by attending to image embeddings, ingredient embeddings, and previously predicted words. Model/Architecture used : The user study results demonstrate the superiority of their system against state-of-the-art image-to-recipe retrieval approaches. (outperforms both human baseline and retrieval based systems obtaining F1 of 49.08%)(good F1 score means that you have low false positives and low false negatives) Model accuracy : They evaluate the whole system on the large-scale Recipe1M dataset Datasets used --------------------------------------------------------- 7. ArcFace: Additive Angular Margin Loss for Deep Face Recognition Full Paper: https://www.catalyzex.com/paper/arxiv:1801.07698 . TLDR: ArcFace can obtain more discriminative deep features and shows state-of-art performance in the MegaFace Challenge in a reproducible way : To enhance intraclass compactness and inter-class discrepancy they propose Additive Angular Margin Loss (ArcFace)- inserting a geodesic distance margin between the sample and centres. This is done to enhance the discriminative power of face recognition model Model/Architecture used : Comprehensive experiments reported demonstrate that ArcFace consistently outperforms the state-of-the-art! Model accuracy : They employ CASIA, VGGFace2, MS1MV2 and DeepGlint-Face (including MS1M-DeepGlint and Asian-DeepGlint) as training data in order to conduct a fair comparison with other methods. Other datasets used- LFW, CFP-FP, AgeDB-30, CPLFW, CALFW, YTF, MegaFace, IJB-B, IJB-C, Trillion-Pairs, iQIYI-VID Datasets used --------------------------------------------------------- 8.Fast Online Object Tracking and Segmentation: A Unifying Approach Full Paper: https://www.catalyzex.com/paper/arxiv:1812.05050 TLDR: The method, dubbed SiamMask, improves the offline training procedure of popular fully-convolutional Siamese approaches for object tracking by augmenting their loss with a binary segmentation task. : SiamMask aims at the intersection between the tasks of visual tracking and video object segmentation to achieve high practical convenience. Like conventional object trackers, it relies on a simple bounding box initialisation and operates online. Differently, from state-of-the-art trackers such as ECO, SiamMask is able to produce binary segmentation masks, which can more accurately describe the target object. SiamMask has 2 variants: three-branch architecture, two-branch architecture (see paper for more details) Model/Architecture used : Qualitative results of SiamMask for both VOT(visual object tracking) and DAVIS (Densely Annotated VIdeo Segmentation) sequences are shown in the paper. Despite the high speed, SiamMask produces accurate segmentation masks even in the presence of distractors. Model accuracy : VOT2016, VOT-2018, DAVIS-2016, DAVIS-2017 and YouTube-VOS Datasets used --------------------------------------------------------- 9. Revealing Scenes by Inverting Structure from Motion Reconstructions Full Paper: https://www.catalyzex.com/paper/arxiv:1904.03303 TLDR:  A team of scientists at Microsoft and academic collaborators reconstruct color images of a scene from the point cloud. :  Our method is based on a cascaded U-Net that takes as input, a 2D multichannel image of the points rendered from a specific viewpoint containing point depth and optionally color and SIFT descriptors and outputs a color image of the scene from that viewpoint. Their network has 3 sub-networks – VISIBNET, COARSENET and REFINENET. The input to their network is a multi-dimensional nD array. The paper explores network variants where the inputs are different subsets of depth, color and SIFT descriptors. The 3 sub-networks have similar architectures. They are U-Nets with encoder and decoder layers with symmetric skip connections. The extra layers at the end of the decoder layers are there to help with high-dimensional inputs Model/Architecture used : The paper demonstrated that surprisingly high quality images can be reconstructed from the limited amount of information stored along with sparse 3D point cloud models Model accuracy :  trained on 700+ indoor and outdoor SfM reconstructions generated from 500k+ multi-view images taken from the NYU2 and MegaDepth datasets Dataset used --------------------------------------------------------- 10. Semantic Image Synthesis with Spatially-Adaptive Normalization Full Paper: https://www.profillic.com/paper/arxiv:1903.07291 TLDR:  Turning Doodles into Stunning, Photorealistic Landscapes! NVIDIA research harnesses generative adversarial networks to create highly realistic scenes. Artists can use paintbrush and paint bucket tools to design their own landscapes with labels like river, rock and cloud Model/Architecture used: In SPADE, the mask is first projected onto an embedding space, and then convolved to produce the modulation parameters γ and β. Unlike prior conditional normalization methods, γ and β are not vectors, but tensors with spatial dimensions. The produced γ and β are multiplied and added to the normalized activation element-wise. In the SPADE generator, each normalization layer uses the segmentation mask to modulate the layer activations. (left) Structure of one residual block with SPADE. (right) The generator contains a series of SPADE residual blocks with upsampling layers. : Our architecture achieves better performance with a smaller number of parameters by removing the downsampling layers of leading image-to-image translation networks. Our method successfully generates realistic images in diverse scenes ranging from animals to sports activities. Model accuracy : COCO-Stuff, ADE20K, Cityscapes, Flickr Landscape Dataset used

Ten Trending Academic Papers on the Future of Computer Vision

What Crypto insiders are reading. There have been a number of media news this week claiming that China’s ban on crypto has been circumvented by VPN and Stablecoins, while folks are going outside of mainland China to do crypto transactions. This is not a surprise to us, and is similar to what we predicted in our previous post when the bans were initially announced. In fact, while we are speaking, the Shanghai Blockchain Week is going on with lots of activities happening. Bitmain CEO Jihan Wu spoke at the conference making predictions for the next decade of blockchain development. He predicts that within a decade, the global blockchain users may exceed 1 billion. Even more interestingly, iQiyi, the Netflix equivalent of China, is airing a live series that follows a crypto enthusiast in China who is looking to survive 21 days in Beijing merely living on 0.21 bitcoin, without any help or donations. The story is also being covered by China Daily, a state-run media outlet. You can see in the iQiyi video that the girl runs around asking all the vendors whether they accept bitcoin as her days are being filmed and streamed. Do you think she’ll survive? Here is a clue: she is now up to day 15. Share with Friends 🌟 Top News in Asia from Monday to Wednesday We have partnered with Hong-Kong based Blockcrunch and 500 startups to sponsor the launch of 500 Blockcrunch, a crypto educational miniseries. The team just released a 30 minute episode covering blockchain scaling, which you can on Apple, Spotify, Overcast and Stitcher. Be sure to check it out! ⁦MaltaStock Exchange ⁩announces MOU with crypto giant ⁦Binance⁩, to set up a security token exchange. http://bit.ly/2p2Trzu *China Media * Caijing, the Chinese financial news site, has acquired the official documents from the court. It ruled that Bitmain’s lawsuit claim on former director of design Yang Zuoxing for patent infringement was rejected. Google translated media | initial story covered by Coindesk Dovey Wan, formerly investment partner at DHVC, and Eric Meltzer, formerly investment partner of INBlock, have just announced their new fund Primitive Ventures. http://bit.ly/2p1TxaH WeChat has blocked the official sales channel of Bitmain. http://bit.ly/2p0jXK8 OKCoin’s founder Star Xu has been released from a police department in Shanghai after 24 hours assisting an investigation into investors’ accusations that OKEx manipulated bitcoin futures on its platform. http://bit.ly/2p2dNJq 💰 Deals and Funds Shanghai blockchain start-up bets on new crypto custody approach despite mainland ban post-funding by Matrix. http://bit.ly/2CI5SdZ Bitmain Co-Lead $14.6M Series A Round In Chinese Video Technology Platform 1000video. http://bit.ly/2oVYQs6 500 Startups raising $5m more for Vietnam-focused fund. http://bit.ly/2p0YZKV 💰 Coin and Token News Former Credit Suisse Managing Director and Head of Corporate Communications Sheel Kohli to Join Block.one as Global Chief Marketing and Communications Officer. http://bit.ly/2oX59fa 💸 Exchange News *China Media* The former director of the Industrial Economic Research Institute of the Ministry of Industry and Information Technology will become the first president of Huobi University. China Media Japan’s internet giant GMO rolls out live crypto trading platform. The group has also established more mining facilities and has mined almost 3,000 BTC and over 660 BCH so far. http://bit.ly/2oVu96C A Crypto trader complained that he lost 20,000 ETH Due to OKEx Malfunction. http://bit.ly/2oZldgE Singapore’s ABCC crypto exchange to actively tap into S. Korean market. http://bit.ly/2oWaovA Japanese team Quoine has unveiled Liquid, a cryptocurrency one-stop trading portal, which will allow users to access a worldwide network of cryptocurrency exchanges. http://bit.ly/2oXxL8b Huobi acquires Singaporean entrepreneur Eric Cheng’s crypto exchange Bit Trade. Bit Trade is a Japanese government-approved platform, which Cheng bought for US$49 million in May 2018. http://bit.ly/2p223Xu 🎌 Regulation News City governments in China have invested $3.57 billion since 2016 to support blockchain startups and R&D. Here’s a guide on who’s doing what. http://bit.ly/2CBQfoe Philippine Digital Asset Exchange is cleared to exchange crypto and the Philippine peso- coins include BTC, ETH, XRP, BCH, and LTC. http://bit.ly/2oYs2Pi Chinese Government accepts blockchain records as evidence in court. http://bit.ly/2oYupl6 S.Korea’s Central Bank continues to sweat over ‘Kimchi Premium’ in cryptocurrency prices. http://bit.ly/2oVtCBE The final court battle for India’s cryptocurrency exchanges begins. http://bit.ly/2p04zgK 💼 Business News Chinese holding company Wuhan General Group, Inc. has entered negotiations to transform a U.S. Defense Department facility into a crypto mining farm. http://bit.ly/2oYROCW Australian Blockchain delegation heads to China to foster fintech ties. http://bit.ly/2p2aemw Chinese Government to use blockchain for tracking charity donations by 2019. http://bit.ly/2p0WisN 🌹🌹🌹 Don’t Forget Check out our guest Interview with Eigen Capital Partner Maomao Hu Joyce Yang from Global Coin Research interviews Maomao Hu, partner at Eigen Capital, a Chinese market maker, and also CTO of Eigen Capital’s sell-side business, Calculus Network, to discuss the evolution of Crypto capital markets in China and the US. Time and Topics discussed in Our Interview 1:18- Introduction to Maomao and Eigen Capital 4:52 — The state of market making in China and the US 11:00- How to hire in China 13:45- Differences between funds in the US and China, is there a difference? 17:00- China’s labor arbitrage 20:17- Chinese Crypto companies’ token incentive structure 26:44- Why Maomao believes that Bitcoin is a threat to the USD 32:01- Is Crypto financial technology? Subscribe

Top Crypto News in Asia- Sept 9- Sept 12

Differences between the tech ecosystem in China vs. the US, from the perspective of a founder that lived in the Bay Area for the last 10 years. Location matters. There’s a reason that talent from across the world congregated in the 50-mile stretch of land spanning San Jose to San Francisco. ( ) In this decade, though, China and its tech centers — Beijing and Shenzhen — are emerging as the Valley’s counter-weight in Asia. Different and full of opportunities. Looking forward 20 years, China will be an epicenter of technology, no doubt about it. You might be interested to check out Steve Blank’s Secret History of Silicon Valley , a wonderful talk on how this place came to be. Personally, my next big decision is where to pursue my next venture — Silicon Valley or China. Through this lens, I embarked on an 8-week trip to China, meeting with friends, entrepreneurs, CEOs, and investors across Beijing, Shanghai, Shenzhen, and Hong Kong. I wanted to learn from their first-hand experience what it’s like to do a business — do a business — in China. Some of these are native-born Chinese; some are Chinese returnees; others are expats. tech These are my findings. It’s not too different. 1) mini-golf on the rooftop of a startup; 2) catered lunch; 3) a startup office park in Shenzhen; 4) a cafe / coworking space in Beijing For a perspective on how it’s like to live in China, check out my previous post: Where Technology Meets Culture: Week 1 of Living in Beijing . China and its businesses are actively embracing technology First, macro trends. Chinese enterprises are wholly embracing technology, across industries. It’s happening across industries, no matter if it’s a traditional or state-owned enterprise, or even how knowledgeable company leadership is about technology. It’s widely accepted that adopting technology — especially artificial intelligence — will “upgrade” the company. Of course, there are gaps in understanding technology across traditional industries and the tech industry, let alone true understanding of AI (which is used more or less synonymously as big data in China). This leads to partnership opportunities, whether by leveraging existing solution such as e-commerce, social networks, or by creating new solutions. There is particular appeal in Silicon Valley technology based on conversations I’ve had with management and investors. Granted, many of these planned partnerships may be shallow and more of a PR stunt in the near-term, but the foresight and determination of a country and its businesses to up-level and willingness to embrace technology cannot be underestimated. The central government is hugely supportive, and AI now plays a prominent part of the government’s current five-year plan ( ). Economist article link Industrial applications per the Fourth Industrial Revolution, just one area where traditional industry is eager to adopt digital technology to “upgrade”. Source: World Economic Forum In the near-term, this presents ample opportunities for creative product managers to bridge business needs with tech solutions. AI is white-hot New York Times posted this in a May 2015 article: ( ) link “It’s a race in the new generation of computing,” said James Lewis, a senior fellow at the Center for Strategic and International Studies. “The difference is that China seems to think it’s a race and America doesn’t.” I think that sums up the attitude in China towards AI. If you think Artificial Intelligence, Machine Learning, Deep Learning are overused words in the valley, try going to a business meeting in Beijing, or even just turn on the TV. True, compared to tech people in the valley, many people probably barely (or not even) understand how it works, but that doesn’t diminish its influence. China is seeing AI as the first tech arena where it can actually compete head-to-head with Silicon Valley. And folks in Beijing may just have an advantage in language. Think about it this way: most engineers in Beijing can easily understand and have access to English-language papers on arxiv. How many engineers in the valley can/will/want to get access to research and discussions made in Chinese in Beijing? In 2017, there were more accepted papers about AI by Chinese authors than American authors. Less than glamorous, but real application. Facial recognition used to dispense toilet paper at public restrooms, so people don’t take the whole roll. Talent, talent, talent Hiring is hard. That’s almost universal. It’s true in China too. First, let’s talk demand. Take this number: 15,000 companies are formed in China. Just think about the demand for talent! every day Then, let’s talk supply. Yes, China has a lot of people and a lot of college graduates, domestic and returnees from study/experience abroad. However, skills vary. While the education system is reforming and graduates are more and more capable, there is still a gap in the average graduate’s ability to work independently and creatively. It takes time to find a candidate. qualified Finally, competition. For a startup to out-compete BAT (Baidu, Alibaba, Tencent), they have to offer the whole package: growth opportunity, equity, higher salaries. Many people simply don’t care about stock options and prefer cash-in-hand. Compared to valley where entrepreneurial engineers will sacrifice salary to join a startup, the culture in China trends conservative. Here’s a glimpse at the comp levels ( ). and Zhihu link All these lead to the fact that competition for talent is fierce. Some of the best talents are poached, oftentimes very aggressively, from other companies. While the best (or most connected) startups are able to build teams full of returnees and top talents, others are less fortunate. Startup culture is driven by a reality of needs To understand startup culture, we need to look deeper at the motivation for the current generation of 20- and 30-somethings. Lecture 1 of YC Startup School ( ) tells, essentially, “don’t do a startup for the money” because joining a “Facebook” in the early stage leads to much better financial outcomes. But the calculus behind that is different in China. YouTube link We need to talk about housing. Let’s do some back of the envelope math. An average “starter home” costs about $1MM both in San Francisco and in Beijing. Beijing, where making a living is a real concern. 1) these nice looking buildings are out of most people’s range. 2) rush hour subway. 3) a , 2br for US$2M condo inside 4th ring In a country where housing prices are stratospherically high, many young adults simply can’t afford a home in their lifetime. Let’s do some simple math. Couple this reality with the rags-to-riches stories of Alibaba, JD, Xiaomi — essentially all the tech giants — that are showcased in spotlight, many entrepreneurial folks decide for themselves that the only way to make it is by starting a company. One may say the economics of it forced their hands. And this is the culture in which the tech ecosystem lives in. And to talk about culture of tech companies, we have to talk about the culture of the people in general. The culture of a company to an extent is dictated by the culture of its customers, and to serve the 1.3 billion local Chinese customers, a company needs to appreciate the local culture. Put in product terms, that’s appreciating the local user behaviors and adapting your product to it, rather than expecting your users to change so they can use your product. When first introduced, Uber was credit card only. In a cash society where only the elite had credit cards (this was pre-mobile payments), Uber was relegated to only the white collar workers and foreigners, while Didi built a strong local base. Differences in customer expectations. Western infrastructure is developed. After all, it’s called “developed” nations for a reason. China’s is not. That presents opportunities to perceptive eyes. But there’s a common string among everyone that capitalizes on those opportunities — they’re ingrained in the local culture. Just as you don’t expect a Chinese student who went study abroad in the US for college to found the next Facebook, you also can’t expect a foreigner, expat, or recent returnee to found the next disruptive company in China. Not alone, at least. The founding team needs to have local *context*. One example I’ve seen is that SaaS companies targeting SMB needs to have a local, offline sales (not customer success) team. What?! Because local consumers expect the guidance for certain services to be done in-person. After all, they’re used to it, because/and labor is so cheap. Blending Silicon Valley and China More and more Chinese nationals are coming back from the US (and in particular from Silicon Valley) to start companies. The companies they found is a blend between the Valley and China. The Valley is often seen as a place of technologists, idealists, and purists. It’s accepted when it comes to hard tech, algorithms, and emerging technology (AI, VR/AR, biotech, etc.), Silicon Valley is the place to go. The Valley is also seen as a place where people are chasing idealistic dreams of “changing the world,” where people do things the proper way. On the other hand, entrepreneurs see China as a place that’s real, down to earth, and practical. People care about the of technology; as a result some of the best product managers grew up in China. Founders and investors care more about cashflow and making a profitable company, and are less patient for pay-off in the long run. application Culture adjustment takes 18 months A founder needs to respect the reality of operating a company in China. On average, founders have told me it took them 1–2 years after they moved from the US to China to fully appreciate and adapt to the macro culture. The best companies run product and engineering like a Silicon Valley company, things like agile, design, collaboration, fail fast, etc. They run operations like a pure blood Chinese company, knowing how to handle government, local partnerships, and veer into the gray areas as needed. Operations is highly efficient If there’s one word I’ve heard over and over that describes running a tech business in China is efficiency. Capital efficiency, time efficiency, resource efficiency. Translated differently, it’s the rallying cry of startups: move fast, and be scrappy. Labor costs are cheaper. This is true both for skilled labor and unskilled labor. A senior engineer at Baidu makes a base of ~US$40k ( ), about a quarter of the costs in Mountain View. If fundraising in the US, that means the same capital now affords you many more engineers. source A popular myth is that local talent isn’t as “good” as US ones, but think about it: China is a country of 1.3 billion people, and a country that highly values education (and that’s an understatement). Many graduates go on to study abroad to get a BS, MS, or PhD. There talents, but just like people anywhere, you have to hire and manage them well. are On-the-ground logistics team of bike sharing companies Having access to cheaper unskilled labor can potentially make an even bigger difference in how a startup operates. Many tasks may not need to be automated at an early stage, but rather handled by people. This allows a company to be more nimble and flexible. It may also directly factor into the business model. To an extent, bike sharing works in China because Mobike/ofo can afford to hire on-the-ground operations team to organize bike parking at every subway station, and shuttle bikes around. 9–9–6 / 9–9–7 This stands for “9am-9pm, 6/7 days a week.” It’s the modus operandi for most early stage startups in China, both founders and early employees. It’s possible because of a generation of young, hungry workers wanting to make something out of their lives. Something magical happens with this round-the-clock cadence. A company moves fast, either iterating quickly towards success, or fails quickly. Either way, no time is wasted; a year feels in China feels like many years elsewhere. Time efficient. 💩 or get off the pot Putting it altogether, it makes the case that building a company in China is actually very efficient. A VC once told me that you can generally tell whether a company is going to make it based on just one month after investment. This is seen time and time again. While a Kickstarter project may be seen as fast if it ships within 6 months, a hardware project is considered late if it doesn’t ship after 2. Each successive round of funding is expected to take place after 6 months. (As an aside, when someone first told me about bike sharing being the hottest thing in China last year, I acknowledged her intellectually but didn’t think much about it. It blew my mind seeing it and trying it in person. Goes to say how much of, literally, legwork is needed to truly understand a market.) SaaS is still in its early days The biggest tech companies in China have all been B2C businesses, BATJ, gaming companies, etc. It makes sense — there’s a huge consumer base and consumption demand. However, it was curious to note the absence of B2B SaaS companies, which constitute a large part of Silicon Valley’s ecosystem. My conclusion after many discussions boils down to this: China hasn’t yet built a habit for buying software. Whereas in the US, people and businesses are happy paying for the value delivered by software, China is coming off two decades of rapid tech adoption marked by rampant piracy. Today’s decision makers aren’t used to paying for a purely digital software without a human touch — SaaS in the vein of Calendly, Gusto, Pipedrive. They didn’t build this habit as consumers, and similarly SaaS never became a trend in the business world. An interesting analogy is in gaming. China is the #1 market for digital games, but it all comes from in-game purchases. People are more willing to buy things in-game (services) vs. the game itself (software). However, this is changing. The youth in this generation is growing up in an environment of digital payments, and are starting to be introduced to subscription models. For example QQ, iQiyi, and Tencent Video all have VIP member options, similar to Spotify and Netflix. This generation’s consumer habits will eventually filter into the workplace. Increasingly, startups are coming to fill this space. A CRM offering Xiaoshouyi has raised over US$50MM. There’s a Google Drive clone in Shimo. Alibaba’s DingDing is a successful Slack clone. It’s only a matter of time before the market of willing-to-pay businesses expands. Conclusion One thing I’ll conclude with. This trip showed me how blinded about China. It’s no longer simply chasing after the US; rather, China and the US have forked in their developments. If you’re at all interested about China, I’d urge you to look out, pay attention, and visit if you can. It’s an exciting world there. Thanks to everyone that I spoke to for their generous sharing of time and insights, and letting me have an inside glimpse into the market. If there’s anything I missed our got wrong, please let me know in a reply. If you like the post, please show your ♥ :)

Stories, News & Mentions about IQIYI

Iqiyi’s Stories on HackerNoon

Iqiyi's News & Mentions