Stayy

stayy

FRESH SUSPICIONS

Intelligent Investing in Cryptocurrencies

Y Combinator reignited the conversation about the role of accelerator programs after publishing some about how applicants that have previously participated in another accelerator are less successful when applying to YC’s programs. interesting findings Ten years in since YC popularized the concept, you’d think that the sheer abundance of “What I learned” post-mortems published on Medium would have been enough to put all but the most effective accelerators out of business. In fact, the opposite has happened: today, it feels as if every university program, every backwoods local chamber of commerce, and every corporate entity is racing to avoid being the last one in on the accelerator game. I used to run an accelerator and spent many sleepless nights trying to figure out how to help each and every one of its portfolio companies equally well (I couldn’t) — I have a major soft spot for these programs and the well meaning people behind them. Still, I worry about the damage many of these newer programs might incur on the greater ecosystem. startup By now, all of the information on how to navigate the early hurdles of starting up is available online. All of this information can be found, answered over and over again, on Quora. All of the advice that you could ever seek out can be found on is being given freely on Medium, or blogs from veteran VC operators like Fred Wilson’s MBA Mondays series, and anything Mark Suster has to say on Both Sides of the Table. So what the hell are you doing paying these accelerator and pre-accelerator programs for basic startup information you can find literally anywhere, for free? Never mind antidilution and polluted cap tables. It’s hard to see straight when you’re eating top ramen for the second month in a row because you’re worth nothing and your company is worth nothing, so let me spell it out for you: even one percent of your company, if you exit at a respectable $200M, is a lot of money. $2M can buy you a lot of nicer ramen. Do you really want to hand that over for, as Andrew Lockley put it, “cheering and a little desk, and little more”? so eloquently With so many accelerators of varying quality competing for startups’ attention and equity, it’s time to take a clear eyed survey of the landscape. Below is a decision tree I hope all founders will use while evaluating whether that startup accelerator program truly makes sense for them. Be most wary of the mediocre actors with good intentions 1) Does it actually offer what you need? Chris Lynch at Accomplice VC caused quite a stir last year when he warned founders to beware of accelerators as . The problem isn’t that these accelerators are intentionally harmful — for every predatory program, there are at least twenty well intentioned others out there taking applications. Unfortunately, as well meaning as they may be, many of the people who run these programs aren’t qualified to do so, whether it’s because they have a limited VC feeder network, have limited entrepreneurial experience, or are out of their depth when it comes to your specific vertical. Have they run their own startup before, cradle to exit? Do they actually have the investor network for you to plug into, once they help you polish that pitch? Don’t be that startup that gets all dressed up with nowhere to go. wolves in sheep’s clothing 2) If yes, are you actually going to get it from the program? Many accelerators are time-constrained by — each cohort gets about eight to twelve weeks of attention, then it’s onto the next shiny batch. Each batch has anywhere from a dozen to even fifty companies. How are you going to guarantee that your 7% will give you access to that top mentor? Would the second- and third- best options even be qualified to effectively help you? They might be if you’re an enterprise SaaS company, but what if you’re in space aquaponics and there’s only mentor knowledgeable in that space — and he gets assigned to someone else? For how to evaluate promises of strategic introductions to customers and distribution partners, see point 4 below. design one 3) If yes, does it make sense right now? We hear the word “accelerator” so often now that we seem to have forgotten the root meaning: accelerators help you move faster. As Paul Graham said in the Hacker News thread about the recent YC findings, “A simple way of thinking about this is that acceleration only helps if you are pointed in the right direction!” Don’t get pushed 100 mph forward only to find yourself having to double back. Most accelerators I’ve seen focus on refining the pitch and pushing you out to investors. I want to see more teach companies how to give investors a strong return on their investment — it seems like beyond coaching pitch and sales tactics, the accelerators leave the lessons on corporate strategy and scaling to the institutional investors. The motivation behind the accelerators’ approach makes sense — for : you’re paying them money to learn how to get more money, which will make more money. This is what accelerators are designed to do, and what they’re best at doing. If you want to take advantage of that, have an MVP and be ready to sell before you go in. them them What about no equity programs? Read beyond the fine print 4) Follow the money: who do they have to answer to? Does that conflict with what you need? Even if they’re not getting it from you, the accelerator has to find money somewhere to keep its lights on and pay its staff. Some of them are sponsored piecemeal by service firms and . The ones that aren’t are still limited in many ways. Dig into their Form 990s and anything else you can find to follow the money and see where it leads. are thinly veiled attempts to sell you recruiting, tax, legal, and other services Local economic development plays Some accelerators rely on government grants and are designed to attract innovation and entrepreneurial talent to an economically blighted area, requiring you to move to a certain city for a specified (and sometimes too long) number of years. They’re often run by former members of chambers of commerce, or local talent with no previous startup experience or startup investor network. (Don’t get me wrong, there are still some fantastic ones that fit this profile but are executed very well, too). And if the region is still recovering economically, will you be able to attract the talent that you need to move quickly? How much time are you going to have to give up to shake hands with visiting politicians and dignitaries? Does it really make sense for your team to stay in Madison, WI for another six months? Or move to Detroit, MI for a year to get that $200,000 grant? As always, it depends. Emerging sector, vertical-specific programs Startup fever is catching on in many creaky old sectors ripe for disruption (fintech) and others that previously crashed and burned but seem primed for a reboot (cleantech, anyone?). Think about it: for sectors with emerging startup activity, the number of people who both know the space well and know how to run a startup is likely small. These people are almost certainly not running your accelerator program — they’re probably chasing startup dreams of their own. For re-emerging sectors that previously experienced a massive flameout like cleantech, the number of people who know what they’re doing is even smaller — otherwise, there would not have been such a massive collapse in the first place. It’s an structural ecosystem problem. Leverage these programs to cut through the noise and grab those strategic partnerships, pilots and customers, but don’t expect much more from most of them. Corporate accelerators Even more are backed by corporations looking for competitive information or to get ahead on the M&A game — would your participation hamper you strategically in the future? Are you locked into using a specific cloud storage service, or certain data visualization platform? Are you prevented from working with certain entities in the future as strategic partners? If so, would it matter to your product development roadmap and growth plans? Look to these above programs for introductions to sources of non-dilutive capital, including cheap loans, free or reduced rent, and grants from governmental and nonprofit agencies. It’s very hard to find and get accepted into a good program. If you don’t get in, don’t waste your time settling for one that isn’t good enough. But what if I’m not in a hot VC market? I count on one hand the number of general-purpose accelerators, and no more than a dozen vertical specific ones, that I’d recommend. I understand the allure of a built-in network of investors when you’re outside one of the current startup hotspots. NewGen recognizes this, and we invest everywhere across the U.S., as do an increasingly large number of VC firms. Trawl Crunchbase and figure out who they are, then reach out. I second Chris Lynch’s declaration: any startup who gets in touch with me will have at least 15 minutes to convince me if I should spend more time learning about them. Seriously — email me with your pitch deck and a 3–4 sentence blurb at ivy@newgenpartners.com. There’s still room for good accelerator programs to be useful. And if your startup is ready for it, knows what you want out of it, and is prepared to fight for it, it can truly accelerate (there’s that word again!) your growth. But just because there’s a wealth of “opportunities” beckoning and offered by the second and worse tier programs, doesn’t mean you have to take it. As an early stage startup, every second of your time and every sub-percentage of your equity counts, because that’s all you have to give. Use it wisely. EDIT: Thanks for the awesome response! Still not sure whether you need an accelerator or should join one right now? Email me with 1) your deck 2) where you’re located, 3) a 2–3 sentence blurb on what you do 4) where you are in your company development, and I’ll help you find an answer. ivy@newgenpartners.com

The startup world has matured. It’s time for most accelerators to go.

Have you ever clicked on a link believing you were heading to one website, only to discover yourself somewhere entirely different? If so, you've had what may be the consequence of an open redirect vulnerability. It may appear to be a little trick or a glitch, but this kind of problem can be an entry point to phishing, malware, and information theft. In this blog, we'll delve into the open redirect vulnerability—what it is, how it works, why it's so risky, and how to avoid it. No technical mumbo-jumbo, no perplexing code—just a simple explanation so you can get a grasp on this security vulnerability. 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/k_gKui-1onY?embedable=true https://youtu.be/k_gKui-1onY?embedable=true What is an Open Redirect? An open redirect is a type of security vulnerability that occurs when a web app permits users to redirect to any external URL without valid verification. from flask import Flask, request, redirect

app = Flask(__name__)

@app.route('/redirect')
def unsafe_redirect():
 target = request.args.get('url')
 return redirect(target) # Open Redirect Vulnerability Here from flask import Flask, request, redirect

app = Flask(__name__)

@app.route('/redirect')
def unsafe_redirect():
 target = request.args.get('url')
 return redirect(target) # Open Redirect Vulnerability Here This type of redirect is commonly caused by a URL parameter. For instance, if you follow a link such as: https://www.openexploit.in/redirect?url=http://verybadwebsite.com https://www.openexploit.in/redirect?url=http://verybadwebsite.com You expect it to take you somewhere safe. But when the website simply accepts any URL fed into that url parameter, attackers can trick users into clicking what seem to be legitimate links but which actually point to something malicious. How Do Attackers Use Open Redirects? Open redirect flaws are frequently misused in phishing attacks. This is how attackers exploit them: Impersonation: The attacker sends a link to a familiar website but with a redirect appended at the end. The user follows the link thinking it's okay—after all, it begins with a trusted domain.
Redirect to Malicious Site: The user then quietly gets redirected to a malicious site. That site may imitate a login page to garner passwords.
Bypassing Security Filters: Certain email filters or firewalls can allow links to trusted domains, even when they have redirect parameters. The attackers take advantage of this by using the good name of the domain as a cover. Impersonation: The attacker sends a link to a familiar website but with a redirect appended at the end. The user follows the link thinking it's okay—after all, it begins with a trusted domain. Impersonation Redirect to Malicious Site: The user then quietly gets redirected to a malicious site. That site may imitate a login page to garner passwords. Redirect to Malicious Site Bypassing Security Filters: Certain email filters or firewalls can allow links to trusted domains, even when they have redirect parameters. The attackers take advantage of this by using the good name of the domain as a cover. Bypassing Security Filters Real-World Example Example 1: Suppose a bank has a site like this: Example 1: https://supersecurebank.com/redirect?url=https://fakesupersecurelogin.com https://supersecurebank.com/redirect?url=https://fakesupersecurelogin.com The bank utilizes this to send customers off to a third-party login service. But the site does not verify whether the url is secure or on an allowlist. # Secure with Allowlist:
from flask import Flask, request, redirect, abort

app = Flask(__name__)

ALLOWED_DOMAINS = ["https://supersecurelogin.com"]

@app.route('/redirect')
def safe_redirect():
 target = request.args.get('url')
 if target in ALLOWED_DOMAINS:
 return redirect(target)
 else:
 abort(400, "Invalid redirect URL.") # Secure with Allowlist:
from flask import Flask, request, redirect, abort

app = Flask(__name__)

ALLOWED_DOMAINS = ["https://supersecurelogin.com"]

@app.route('/redirect')
def safe_redirect():
 target = request.args.get('url')
 if target in ALLOWED_DOMAINS:
 return redirect(target)
 else:
 abort(400, "Invalid redirect URL.") Example 2: A crook sends an email: Example 2: "Urgent: Your account is suspended. Click here to verify your identity upon login." The majority of the users think that the link begins with supersecurebank.com and, therefore, are confident. However, as they click the link, they're immediately routed to a convincing pretend login web page—and this is where credentials get stolen. How Do Open Redirects Occur? Open redirects tend to happen because of: Lack of validation: The website doesn't validate if the redirect URL is safe or authorized.


Too many URL parameters: Sites employ parameters such as ?redirect=, ?url=, or ?next= to manage navigation but do not limit rules about what values are allowed.


User input trust: Using whatever the user submits in a URL without checking is inherently dangerous.
Vulnerable Example
const express = require('express'); 
const app = express();
app.get('/redirect', (req, res) => { const url = req.query.url; res.redirect(url); // Vulnerable to open redirect });

Secure Example
const express = require('express');
const app = express();

const allowedDomains = ['https://openexploit.in', 'https://www.openexploit.in'];

app.get('/redirect', (req, res) => { const target = req.query.url;

if (allowedDomains.includes(target)) { 
 res.redirect(target); } 
else { res.status(400).send('Invalid redirect URL.'); 
} 
}); Lack of validation: The website doesn't validate if the redirect URL is safe or authorized. Lack of validation: The website doesn't validate if the redirect URL is safe or authorized. Too many URL parameters: Sites employ parameters such as ?redirect=, ?url=, or ?next= to manage navigation but do not limit rules about what values are allowed. Too many URL parameters: Sites employ parameters such as ?redirect=, ?url=, or ?next= to manage navigation but do not limit rules about what values are allowed. User input trust: Using whatever the user submits in a URL without checking is inherently dangerous.
Vulnerable Example
const express = require('express'); 
const app = express();
app.get('/redirect', (req, res) => { const url = req.query.url; res.redirect(url); // Vulnerable to open redirect });

Secure Example
const express = require('express');
const app = express();

const allowedDomains = ['https://openexploit.in', 'https://www.openexploit.in'];

app.get('/redirect', (req, res) => { const target = req.query.url;

if (allowedDomains.includes(target)) { 
 res.redirect(target); } 
else { res.status(400).send('Invalid redirect URL.'); 
} 
}); User input trust: Using whatever the user submits in a URL without checking is inherently dangerous. Vulnerable Example const express = require('express'); 
const app = express();
app.get('/redirect', (req, res) => { const url = req.query.url; res.redirect(url); // Vulnerable to open redirect }); const express = require('express'); 
const app = express();
app.get('/redirect', (req, res) => { const url = req.query.url; res.redirect(url); // Vulnerable to open redirect }); Secure Example const express = require('express');
const app = express();

const allowedDomains = ['https://openexploit.in', 'https://www.openexploit.in'];

app.get('/redirect', (req, res) => { const target = req.query.url;

if (allowedDomains.includes(target)) { 
 res.redirect(target); } 
else { res.status(400).send('Invalid redirect URL.'); 
} 
}); const express = require('express');
const app = express();

const allowedDomains = ['https://openexploit.in', 'https://www.openexploit.in'];

app.get('/redirect', (req, res) => { const target = req.query.url;

if (allowedDomains.includes(target)) { 
 res.redirect(target); } 
else { res.status(400).send('Invalid redirect URL.'); 
} 
}); Is It a Common Vulnerability? Yes, extremely. Open redirect problems have cropped up even in big organizations such as Google, Facebook, and Microsoft previously (though they were rapidly addressed). It's also categorized in OWASP's Unvalidated Redirects and Forwards, a recognized category of vulnerabilities developers ought to be on the lookout for. How to Prevent Open Redirects? Prevention involves cautious management of redirects within your web application. Below are some safe practices: Avoid Open Redirects Altogether

If you don't absolutely have to redirect users depending on URL parameters, don't. Less is more when it comes to design.



Implement a Redirect Allowlist

Permit redirects only to domains or URLs that are explicitly known to be trusted. If a user attempts to redirect somewhere else, deny the request.

Secure Coding in Python
@app.route('/redirect') 
def relative_redirect(): 
next_page = request.args.get('next', '/') # Only allow internal redirects 
if next_page.startswith('/'): 
 return redirect(next_page) 
else: 
 abort(400, "External redirects are not allowed.")

Secure Coding in Express
app.get('/redirect', (req, res) => { 
const next = req.query.next; if (next && next.startsWith('/')) { 
res.redirect(next); } else { res.status(400).send('Invalid redirect path'); } });



Validate User Input

Sanitize and check any redirect URLs. For instance:
Ensure the URL is relative (internal only), not a full external link.
Block "http://" or "https://" URLs unless they're on a safe list.



Warn Users Before Redirecting


Show a message such as: "You are leaving securebank.com and traveling to another site: fakebank-login.com. Do you continue?"


This makes users more careful about what they're leaving for.
@app.route('/redirect') 
def warn_redirect(): 
url = request.args.get('url') 
return f""" <html> <body> You are being redirected to: {url} <a href="{url}">Click here if you trust the destination</a> </body> </html> """





Log and Monitor

Monitor how frequently redirects occur and where they go. Unusual patterns might indicate abuse. Avoid Open Redirects Altogether

If you don't absolutely have to redirect users depending on URL parameters, don't. Less is more when it comes to design. Avoid Open Redirects Altogether If you don't absolutely have to redirect users depending on URL parameters, don't. Less is more when it comes to design. If you don't absolutely have to redirect users depending on URL parameters, don't. Less is more when it comes to design. Implement a Redirect Allowlist

Permit redirects only to domains or URLs that are explicitly known to be trusted. If a user attempts to redirect somewhere else, deny the request.

Secure Coding in Python
@app.route('/redirect') 
def relative_redirect(): 
next_page = request.args.get('next', '/') # Only allow internal redirects 
if next_page.startswith('/'): 
 return redirect(next_page) 
else: 
 abort(400, "External redirects are not allowed.")

Secure Coding in Express
app.get('/redirect', (req, res) => { 
const next = req.query.next; if (next && next.startsWith('/')) { 
res.redirect(next); } else { res.status(400).send('Invalid redirect path'); } }); Implement a Redirect Allowlist Permit redirects only to domains or URLs that are explicitly known to be trusted. If a user attempts to redirect somewhere else, deny the request. Permit redirects only to domains or URLs that are explicitly known to be trusted. If a user attempts to redirect somewhere else, deny the request. Secure Coding in Python @app.route('/redirect') 
def relative_redirect(): 
next_page = request.args.get('next', '/') # Only allow internal redirects 
if next_page.startswith('/'): 
 return redirect(next_page) 
else: 
 abort(400, "External redirects are not allowed.") @app.route('/redirect') 
def relative_redirect(): 
next_page = request.args.get('next', '/') # Only allow internal redirects 
if next_page.startswith('/'): 
 return redirect(next_page) 
else: 
 abort(400, "External redirects are not allowed.") Secure Coding in Express app.get('/redirect', (req, res) => { 
const next = req.query.next; if (next && next.startsWith('/')) { 
res.redirect(next); } else { res.status(400).send('Invalid redirect path'); } }); app.get('/redirect', (req, res) => { 
const next = req.query.next; if (next && next.startsWith('/')) { 
res.redirect(next); } else { res.status(400).send('Invalid redirect path'); } }); Validate User Input

Sanitize and check any redirect URLs. For instance:
Ensure the URL is relative (internal only), not a full external link.
Block "http://" or "https://" URLs unless they're on a safe list. Validate User Input Sanitize and check any redirect URLs. For instance:
Ensure the URL is relative (internal only), not a full external link.
Block "http://" or "https://" URLs unless they're on a safe list. Sanitize and check any redirect URLs. For instance: Ensure the URL is relative (internal only), not a full external link. Block "http://" or "https://" URLs unless they're on a safe list. Warn Users Before Redirecting


Show a message such as: "You are leaving securebank.com and traveling to another site: fakebank-login.com. Do you continue?"


This makes users more careful about what they're leaving for.
@app.route('/redirect') 
def warn_redirect(): 
url = request.args.get('url') 
return f""" <html> <body> You are being redirected to: {url} <a href="{url}">Click here if you trust the destination</a> </body> </html> """ Warn Users Before Redirecting Show a message such as: "You are leaving securebank.com and traveling to another site: fakebank-login.com. Do you continue?"


This makes users more careful about what they're leaving for.
@app.route('/redirect') 
def warn_redirect(): 
url = request.args.get('url') 
return f""" <html> <body> You are being redirected to: {url} <a href="{url}">Click here if you trust the destination</a> </body> </html> """ Show a message such as: "You are leaving securebank.com and traveling to another site: fakebank-login.com. Do you continue?" Show a message such as: "You are leaving securebank.com and traveling to another site: fakebank-login.com. Do you continue?" This makes users more careful about what they're leaving for.
@app.route('/redirect') 
def warn_redirect(): 
url = request.args.get('url') 
return f""" <html> <body> You are being redirected to: {url} <a href="{url}">Click here if you trust the destination</a> </body> </html> """ This makes users more careful about what they're leaving for. @app.route('/redirect') 
def warn_redirect(): 
url = request.args.get('url') 
return f""" <html> <body> You are being redirected to: {url} <a href="{url}">Click here if you trust the destination</a> </body> </html> """ @app.route('/redirect') 
def warn_redirect(): 
url = request.args.get('url') 
return f""" <html> <body> You are being redirected to: {url} <a href="{url}">Click here if you trust the destination</a> </body> </html> """ Log and Monitor

Monitor how frequently redirects occur and where they go. Unusual patterns might indicate abuse. Log and Monitor Monitor how frequently redirects occur and where they go. Unusual patterns might indicate abuse. Monitor how frequently redirects occur and where they go. Unusual patterns might indicate abuse. What Should Users Do to Stay Safe? Even if a site contains an open redirect, visitors can safeguard themselves: Hover over links prior to clicking to find out where they point.
Don't believe URLs at face value—just because it begins with a familiar site doesn't make it secure.
Employ browser protection and security add-ons that notify you of phishing sites.
Watch out for desperate emails or messages requiring you to log in quickly or confirm your information. Hover over links prior to clicking to find out where they point. Don't believe URLs at face value—just because it begins with a familiar site doesn't make it secure. Employ browser protection and security add-ons that notify you of phishing sites. Watch out for desperate emails or messages requiring you to log in quickly or confirm your information. Conclusion Whether you're a web developer, security engineer, or simply a curious internet browser, knowing how open redirects are used and why they're a problem is a critical part of being safe on the web. Keep in mind: The best defense is a good understanding. Now that you know what an open redirect is, you'll be more effective at staying clear of its risks and keeping others safe as well.

A Good Understanding of Open Redirect Vulnerability Can Help You Stay Safe Online

It’s 2025, and AI has officially checked into every part of healthcare—from the front desk scheduler to the operating room. It reminds your patient to refill their medication and helps flag cancer on a slide. It doesn’t complain, take breaks or call in sick. It’s fast, tireless and… maybe not enough? Because for all its progress, we’re still seeing AI systems stumble in real clinical settings. It might not handle edge cases as effectively as we’d like. It misses context that experienced clinicians could catch before their first cup of coffee. And no, this isn’t about fearing automation or fighting the march of progress. I’ve spent years building and implementing these systems, and I know firsthand that AI can be an incredibly effective collaborator… but only when it’s built to work with people. with It might be true that in some industries, you can plug AI in and get pretty incredible outcomes, with a few duds here and there. But healthcare doesn’t have much margin for error. What matters is good judgment and context, and that takes human input. Let’s talk about what it means to build AI infrastructure that puts humans in the loop by design. AI Isn’t Failing Because It’s Wrong—It Just Needs Help Starting with data, because it’s where your AI will: AI models want structure. Healthcare gives them chaos. Any clinician is familiar with the digital mess behind the scenes, ranging from poorly scanned PDFs of half-filled forms to different EHRs calling the same condition by different names. An algorithm might see an “abnormal” lab and sound the alarm, not realizing the patient just had surgery last week. Or maybe there’s a coffee stain over a 0. According to ECRI’s 2025 safety report, AI-driven clinical errors, ranging from misdiagnoses to inappropriate treatment recommendations, are now considered one of the top patient risks. And to be clear, the models aren’t malicious. But letting them fly blind and build a foundation on human error is asking for trouble. We need to provide a scaffolding for uncertainty and ambiguity. And coffee stains. ECRI’s 2025 safety report ECRI’s 2025 safety report This is a huge context problem, and wouldn’t you know it: humans are very good at context. AI sees the signal. Humans know the story. The best results happen when AI proposes and humans decide. False Positives Create More Work Let’s follow that same example further. What happens when AI goes off on its own? A radiology AI cries wolf and flags nodules of interest in every single scan, including the ones that are just shadows or scar tissue. Trust falls off, teams start to ignore it. Then usage drops. At some point, the system gets shelved. So much for a decision support system. decision support system This is a classic tech-in-medicine problem. A 2023 NPJ Digital Medicine study flagged this exact pattern. AI without oversight created so many false positives that it added to staff burden, increasing alert fatigue and driving its users to disengage. 2023 NPJ Digital Medicine study 2023 NPJ Digital Medicine study We’ve seen in the field again and again that a model can’t triage itself, and it certainly won’t know whether its prediction is helpful or harmful. That’s the role of a human reviewer: someone needs to be partnered and trained to verify and decide on what action needs to happen next. It’s an important layer of checks and balances that should drive every AI-assisted medical decision. Black-Box Models Don’t Fly in Healthcare This shouldn’t come as a surprise: Even when the model is right, doctors need to know why. If your AI tells a provider that patient X is at high risk of deterioration, they’re going to ask: based on what evidence? Is it recent vitals? A pattern? If the model can’t explain itself, or if it dumps raw scores with no context, it’s not providing meaningful or actionable output. A 2024 American Medical Association survey found that nearly half of physicians rank transparency and human oversight as top criteria for trusting an AI tool. Accuracy is one thing, but someone has to be accountable, and it will always be the practicing physician. 2024 American Medical Association survey 2024 American Medical Association survey That’s where human-in-the-loop design really counts. Reviewers vet the alert, while clinical leads can escalate or override. Everyone sees what changed and why, and knows who to turn to for questions. Introducing HITL turns your AI from a black box into a system with feedback and contextual memory. Regulations Are Catching Up The FDA, WHO, and pretty much every major regulatory body now agrees: AI in healthcare needs oversight, even in day-to-day ops. The FDA’s 2025 guidance emphasizes traceability and documented human oversight at each major decision. The World Health Organization warns against deploying AI without rigorous evaluation and expert supervision. A very recent U.S. Senate investigation found that fully autonomous AI systems in insurance led to care denials 16 times higher than expected, calling for immediate policy intervention. I won’t mince words: If you’re not building with humans in the loop, you’re inviting much more scrutiny in an already audit-heavy industry. But Humans Are Slow... Right? Not If You Design for It There’s that persistent myth that adding people into an AI system slows things down. But that’s only true if the system wasn’t designed to support collaboration in the first place. A few ways to do it without bogging down your ops: Set confidence thresholds. Only flag low-confidence cases, where human input adds value.
Build effective tooling. Reviewers need intuitive and actionable dashboards, not a 50-column CSV.
Track outcomes. Log every review, override and resolution. You’ll be ready for the next audit and simultaneously build a goldmine of data for model improvement. Set confidence thresholds. Only flag low-confidence cases, where human input adds value. Build effective tooling. Reviewers need intuitive and actionable dashboards, not a 50-column CSV. Track outcomes. Log every review, override and resolution. You’ll be ready for the next audit and simultaneously build a goldmine of data for model improvement. It’s easy to start thinking it’s bureaucratic overhead, but a well-designed HITL system becomes a learning system. The more it runs, the better it gets. If You Want to Scale AI in Healthcare, Start With Humans Let’s recap. We don’t need to guess where things go wrong. AI models are left in a vacuum, trusted to succeed with poor inputs, then start eroding trust through noisy outputs. Clinicians tune out and the systems get shut off. But healthcare has never been about hitting 99% accuracy in a sandbox. We focus on making the right call when everything’s messy to begin with. That’s what makes human judgment irreplaceable, and why we have the responsibility to design with the people who still carry the weight of patient care. By all means, let AI handle the routine stuff. Automate the admin, surface insights, reduce friction where it makes sense. But when it comes to escalation and practicing good judgment, build systems that know when to hand it back to a human. In the future, we might start realizing the smartest thing an AI can do is know when to ask for help.

AI in Healthcare Works—But Only When Humans Stay in the Loop

Authors:
(1) Zhenbang Feng, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (jasonfen@usc.edu);
(2) Hardhik Mohanty, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (hmohanty@usc.edu);
(3) Bhaskar Krishnamachari, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (bkrishna@usc.edu). Authors: Authors: (1) Zhenbang Feng, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (jasonfen@usc.edu); (2) Hardhik Mohanty, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (hmohanty@usc.edu); (3) Bhaskar Krishnamachari, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (bkrishna@usc.edu). Editor’s note: This article is part of a broader study. You’re reading part 4 of 5. Read the rest below. Editor’s note: This article is part of a broader study. You’re reading part 4 of 5. Read the rest below. Editor’s note: This article is part of a broader study. You’re reading part 4 of 5. Read the rest below. Table of Links Abstract and I. Introduction
II. The Mechanism
III. Price-Stabilizing Mechanisms, IV. Simulating DAI & V. Simplified Mathematical Model for DAI
VI. How Collateral Price Impacts Derivative Price
VII. Risk Factors for Stable Derivatives & VIII. Conclusions and References Abstract and I. Introduction Abstract and I. Introduction II. The Mechanism II. The Mechanism III. Price-Stabilizing Mechanisms, IV. Simulating DAI & V. Simplified Mathematical Model for DAI III. Price-Stabilizing Mechanisms, IV. Simulating DAI & V. Simplified Mathematical Model for DAI VI. How Collateral Price Impacts Derivative Price VI. How Collateral Price Impacts Derivative Price VII. Risk Factors for Stable Derivatives & VIII. Conclusions and References VII. Risk Factors for Stable Derivatives & VIII. Conclusions and References VI. HOW COLLATERAL PRICE IMPACTS DERIVATIVE PRICE A. Empirical evidence A. Empirical evidence Figure-2 demonstrates the closing prices of both ETH and DAI over the two-year period from 2018 to 2020. Ethereum (ETH) exhibited significant volatility, with its price ranging from approximately $84 to $1456. The average closing price for ETH during this period was about $332.52. On the other hand, DAI demonstrated its characteristic stability. As a stablecoin, its price hovered around the $1 peg, with minor fluctuations. The average closing price for DAI was approximately $1.0007. To further understand the relationship between ETH and DAI, we calculated the correlation coefficient between their closing prices, which came out as 0.1336. The calculated correlation coefficient of 0.1336 suggests a weak positive linear relationship between the closing prices of ETH and DAI. The correlation coefficient being close to 0 is weak enough to draw any meaningful conclusions. Figure-3 illustrates the scatter plot of DAI vs ETH closing prices. We observe there’s a high density of points around the DAI price of $1, confirming its stable nature. This dense cluster is spread across various ETH prices, which is consistent with the observation that DAI maintains its peg to around $1 regardless of ETH’s price movements. The scatter plot reaffirms the distinct characteristics of the two assets. DAI, being a stablecoin, predominantly hovers around the $1 mark, irrespective of the volatility and price changes exhibited by ETH. This scatter plot also aligns with the previously computed weak correlation coefficient, indicating that the movements in DAI prices have a weak linear relationship with ETH’s price movements. B. Simulation results B. Simulation results Figures-4 & 6 demonstrates the predicted value of the price of DAI through simulation models. The DAISIM simulation model initially outputs DAI prices ranging from approximately $3.34 to $4.17. This shows that the predicted DAI price significantly deviates from its pegged value of $1. This variation is attributed to the model’s lack of consideration for market beliefs and narratives. However, with the integration of a belief parameter based on the hypothesis of DAI’s peg to $1, there is a substantial improvement in price stability. The modified model generates DAI prices that fluctuate closely around the $1 mark with minimal deviations. This improvement is also evidenced by examining the correlation coefficients between DAI and ETH prices in both models as depicted in Figures-5 & 7. Initially, the correlation coefficient stands at 0.86, indicating a strong positive correlation. This suggests that DAI prices are significantly influenced by ETH’s price movements, contrary to the observed behavior of DAI in the real world. However, after the introduction of the belief parameter, the correlation coefficient decreases to 0.54. The lower correlation more accurately mirrors the real-world expectation that DAI’s price remains largely independent of ETH’s market fluctuations. This observation reinforces the significance of integrating realistic market factors into financial simulation models. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

DAI Stays at $1 Even When ETH Goes Crazy. Here’s the Data.

Authors:
(1) Zhenbang Feng, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (jasonfen@usc.edu);
(2) Hardhik Mohanty, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (hmohanty@usc.edu);
(3) Bhaskar Krishnamachari, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (bkrishna@usc.edu). Authors: Authors: (1) Zhenbang Feng, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (jasonfen@usc.edu); (2) Hardhik Mohanty, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (hmohanty@usc.edu); (3) Bhaskar Krishnamachari, Department of Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern, California Los Angeles, CA, USA (bkrishna@usc.edu). Editor’s note: This article is part of a broader study. You’re reading part 3 of 5. Read the rest below. Editor’s note: This article is part of a broader study. You’re reading part 3 of 5. Read the rest below. Editor’s note: This article is part of a broader study. You’re reading part 3 of 5. Read the rest below. Table of Links Abstract and I. Introduction
II. The Mechanism
III. Price-Stabilizing Mechanisms, IV. Simulating DAI & V. Simplified Mathematical Model for DAI
VI. How Collateral Price Impacts Derivative Price
VII. Risk Factors for Stable Derivatives & VIII. Conclusions and References Abstract and I. Introduction Abstract and I. Introduction II. The Mechanism II. The Mechanism III. Price-Stabilizing Mechanisms, IV. Simulating DAI & V. Simplified Mathematical Model for DAI III. Price-Stabilizing Mechanisms, IV. Simulating DAI & V. Simplified Mathematical Model for DAI VI. How Collateral Price Impacts Derivative Price VI. How Collateral Price Impacts Derivative Price VII. Risk Factors for Stable Derivatives & VIII. Conclusions and References VII. Risk Factors for Stable Derivatives & VIII. Conclusions and References III. PRICE-STABILIZING MECHANISMS Synthetic derivatives like Single-Collateral DAI (SAI) aim to mirror a particular asset’s value. In SAI’s case, it seeks to peg its value to the US Dollar. The question then arises: How does it ensure that its value remains stable at 1 USD? In reality, several mechanisms work together to ensure this stability. A. Emergency Shutdown A. Emergency Shutdown Emergency shutdown is a mechanism designed for extreme circumstances. This process can be triggered during events like security breaches or huge market volatility. It allows DAI holders to directly redeem its value for $1 of collateral at the point in time when the Emergency Shutdown is initiated, essentially enforcing the 1:1 USD soft peg. B. DAI Savings Rate or Stability Fees B. DAI Savings Rate or Stability Fees Another significant mechanism to ensure price stability is the DAI Savings Rate. When market dynamics push the DAI price away from the target, this savings rate is adjusted. If DAI’s market price goes above 1 USD, the savings rate decreases, reducing demand and pushing the price towards 1 USD. Conversely, if DAI’s price falls below 1 USD, the savings rate rises, increasing demand and nudging the price upwards. In the Single-Collateral DAI, there isn’t a DAI Savings Rate; instead, stability Fees on CDPs are adjusted directly to balance the demand and supply. C. Keepers C. Keepers Keepers play an invaluable role in maintaining DAI’s stability. These are typically automated agents incentivized by profit. They actively participate in Debt and Collateral Auctions during CDP liquidations. More crucially, they trade DAI around its target price, selling when the price is high and buying when it’s low, banking on the long-term convergence towards the target price, thereby helping in maintaining the price stability. Individuals or entities are often motivated to operate keepers due to the potential for profit. This profit expectation relies on their belief that the price of DAI will remain stable around 1 USD. D. Narrative or Belief D. Narrative or Belief An often underestimated element in the mechanism of stability is the narrative or the shared belief of its participants. While tangible mechanisms like target prices, keepers, and oracles play vital roles, the intangible narrative equally underpins the stability of such systems. For example, the shared belief in DAI(SAI) is that it will stay aligned with the US dollar. This belief is propagated and reinforced through consistent communication, historical performance, community trust, and governance transparency. When participants uniformly believe in the 1:1 pegging, they act in ways that naturally enforce this peg further. For instance, if the price of SAI drifts slightly above 1 USD, the narrative would drive participants who are believing to sell, thereby pushing the price back to the peg. Conversely, if SAI dips below 1 USD, the narrative would deem it undervalued, triggering buys and bringing the price back to one dollar. This shared narrative isn’t solely built on blind faith. When the system consistently demonstrates its ability to maintain the peg through various market conditions, it reinforces the narrative. In the following sections, we will delve deeper into the role of this narrative by simulating the effects of shared belief. Role of Belief on DAI price dynamics: Let’s consider the DAI to be solely influenced by “belief” (when b is high). It becomes apparent that a shared consensus on price among the majority of investors creates a situation where an individual investor with a different view is likely to experience a negative expected return. Thus, an investor without extra information would tend to follow the majority, which again reinforces the consensus. Let’s consider two illustrative scenarios: Role of Belief on DAI price dynamics: Let’s consider the DAI to be solely influenced by “belief” (when b is high). It becomes apparent that a shared consensus on price among the majority of investors creates a situation where an individual investor with a different view is likely to experience a negative expected return. Thus, an investor without extra information would tend to follow the majority, which again reinforces the consensus. Let’s consider two illustrative scenarios: • First, suppose the current price of DAI is 1.2 USD. If investor Bob believes that the price will rise to 1.5 USD, he would purchase or take a long position in DAI, while the majority, anticipating a price decline, would sell or short DAI. • Second, consider a situation where the price of DAI is 0.8 USD. If Bob expects the price to drop further to 0.5 USD, he would choose to sell or short DAI. However, if the majority believes the price will increase, they would be looking to buy or go long on DAI. In essence, any given price could theoretically serve as a Nash Equilibrium. However, the mechanism of over-collateralization, Emergency Shutdown, and narrative set by MakerDAO suggest that the Schelling point is anchored at 1 USD. IV. SIMULATING DAI A. DAISIM A. DAISIM DAISIM [12] presents a detailed computational simulation of the single-collateral DAI stablecoin, which was initiated by the MakerDAO project in 2017. This study places a significant emphasis on modeling the behavior of cryptocurrency investors, who are characterized by a range of risk tolerances and are considered rational in their portfolio optimization strategies. A key feature of this simulator is its incorporation of automated mechanisms for both order matching and updating prices, which are critical for determining the market price of DAI. The primary aim of this research is to investigate how the price of single-collateral DAI and the allocation of investment portfolios vary among a group of investors. This variation is analyzed in the context of exogenous parameters such as the price of ETH, as well as several internal system parameters, including the stability rate and transaction fees. B. Updated objective function B. Updated objective function However, the simulated price of the DAI driven by this objective function displayed significant deviation from the desired 1 USD peg. Such consistent deviations suggest that external factors, potentially rooted in market beliefs about DAI’s stability were absent from the model’s considerations. Therefore, we introduce “belief factor” (β) to better encapsulate market sentiment regarding DAI’s valuation. Incorporating this belief factor accounts for the dual nature of market reactions to DAI’s price fluctuations. When the price of DAI falls below $1, the positive β terms stimulate an increase in demand and a decrease in supply, aiming to correct the price towards its intended $1 peg. Conversely, when DAI’s price exceeds $1, these terms adjust to increase supply and reduce demand, facilitating a return to the pegged price. When integrated into the model dynamics, these adjustment mechanisms highlight the investor’s expectation that DAI should maintain a value of $1. Hence, the refined objective function to be maximized can be written in the following manner: V. SIMPLIFIED MATHEMATICAL MODEL FOR DAI Here we present a simple theoretical model of DAI price. This model is presented primarily as an explanatory tool - we do not claim that this is the “correct” model of reality. However, this model allows us to explain two phenomena: 1) how belief influences the price being close to 1, reflected in a high demand for DAI at price below that value and a low demand beyond that price, and 2) how dependence of supply and demand for DAI on the price of ETH might impact the relationship between ETH price and the valuation of DAI. This model could potentially serve as a baseline for further investigation and development by the research community. To determine the price, we equate these two as follows: This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

DAI Doesn’t Stay at $1 by Accident—Here’s the System That Keeps It There

Web scraping streamlines data collection from websites, aiding in competitor analysis, content migration, and machine learning training. While scraping public data is legal, businesses continue to tighten restrictions. So, is it really a new year, new rules? In this article, we’ll break it down and share practical tips to keep you ahead. new year, new rules What’s on the Horizon? What’s on the Horizon? AI is no secret - its rapid development has led to smarter, more adaptive anti-scraping systems. As AI advances, anti-bot systems are becoming more aggressive. The result? Scrapers often come across unexpected roadblocks. Bot detection powered by AI analyzes patterns and behaviors that set bots apart from human users, including factors like IP address activity, browsing speed, mouse movements, and other signals. AI algorithms can adapt to the techniques bots use. As a result, even when scrapers change tactics - such as simulating human actions - the AI can still spot unusual patterns, triggering CAPTCHAs or blocking the request altogether. Moreover, AI can analyze a wider range of data sources to identify anomalies, like frequent requests from the same IP in a short time, which is a common sign of scraping. Another aspect is advanced CAPTCHAs. They combine multiple layers of defense, from object recognition and behavioral analysis to machine learning. Some systems mix CAPTCHAs with two-factor authentication (2FA). It means that even if a bot bypasses a CAPTCHA challenge, additional layers of security will block it. Some websites may integrate biometric verification or use cryptographic puzzles. Next, it’s a technique that intentionally makes JavaScript code difficult to read. JavaScript-based obfuscation is a method where the code written in JavaScript is altered to make it more complex, which makes it difficult for scrapers to extract data. In 2025 and beyond, obfuscation may become part of a more comprehensive anti-scraping strategy, combining with CAPTCHAs, AI-based bot detection, and behavioral analysis to create a multi-layered defense against automated scraping. If you're a developer working with a company providing access to financial data, you'll need an API key, which will authenticate the request and ensure valid access. OAuth and API keys are two common authentication methods. When a user logs into an app via their Google or Facebook account, OAuth is used to grant permission to the app to access their profile information or social media data without sharing their password. These methods will remain key for businesses to secure data and maintain user privacy, all while supporting third-party developer partnerships. In 2025, platforms will use advanced fingerprinting and IP blocking to prevent scraping. Services like Netflix, Hulu, and BBC iPlayer will continue to employ geo-blocking, making it harder for scrapers to access restricted content. To bypass these, rotating proxies and VPNs will be a must-have but more challenging to manage. What’s on the Legal Side? What’s on the Legal Side? One of the most significant factors affecting the future of web scraping is the increase in data privacy laws worldwide. Governments are tightening regulations on the collection, processing, and storage of personal data. The General Data Protection Regulation (GDPR) in the European Union, for example, requires organizations to obtain explicit consent before collecting personal data. This has significantly impacted web scraping, particularly for websites dealing with personal information. Furthermore, Terms of Service (ToS) agreements are increasingly rigorous, as numerous platforms clearly ban scraping activities. Companies like Amazon, Google, and eBay have taken legal action against violators, enforcing strict rules around scraping product listings, reviews, and exclusive data. Consequently, many companies are prioritizing the compliance of third-party data sources with both local and international laws. You might think that scraping without legal permission could offer short-term benefits, but it's important to think twice. Always review the Terms of Service and ensure your actions are in compliance. It's better to be safe than sorry. Always review the Terms of Service and ensure your actions are in compliance. Put an End to Scraping Headaches Put an End to Scraping Headaches As you can see, the rules are constantly evolving, but for any professional, this shouldn’t be a problem. Here are some smart strategies to manage the growing difficulties in web scraping. First things first - rotating residential proxies. They are unique IPs of real devices. Websites rarely detect such proxies, so users can scrape data without blocks. Traffic appears natural when routed through residential proxies. At DataImpulse, you can also test Premium residential proxies with full customization options.
Use AI-powered CAPTCHA solvers. These tools incorporate advanced algorithms to crack complex puzzles. Overall, AI solvers can handle CAPTCHAs by using improved recognition and machine learning capabilities.
Encrypt your traffic. Consider using TLS for secure connections and use HTTP/2 headers to speed up your requests.
If possible, use official APIs to reduce legal risks and minimize detection instead of traditional scraping.
Set randomized request intervals. Mimic human browsing behavior by setting random intervals between requests.
Choose distributed scraping for large-scale tasks. Run scrapers on multiple cloud servers or edge computing nodes to balance traffic and reduce suspicion.
Use ISP and mobile proxies. They provide better anonymity and are less likely to get blocked (compared to datacenter proxies).
Implement fingerprint spoofing. Modify browser fingerprints (user agents, canvas, WebGL, etc.)
Work with legal experts. You can have a consultation with legal professionals to understand the possible outcomes of web scraping. First things first - rotating residential proxies. They are unique IPs of real devices. Websites rarely detect such proxies, so users can scrape data without blocks. Traffic appears natural when routed through residential proxies. At DataImpulse, you can also test Premium residential proxies with full customization options. residential residential Premium Premium Use AI-powered CAPTCHA solvers. These tools incorporate advanced algorithms to crack complex puzzles. Overall, AI solvers can handle CAPTCHAs by using improved recognition and machine learning capabilities. Encrypt your traffic. Consider using TLS for secure connections and use HTTP/2 headers to speed up your requests. If possible, use official APIs to reduce legal risks and minimize detection instead of traditional scraping. Set randomized request intervals. Mimic human browsing behavior by setting random intervals between requests. Choose distributed scraping for large-scale tasks. Run scrapers on multiple cloud servers or edge computing nodes to balance traffic and reduce suspicion. Use ISP and mobile proxies. They provide better anonymity and are less likely to get blocked (compared to datacenter proxies). Implement fingerprint spoofing. Modify browser fingerprints (user agents, canvas, WebGL, etc.) Work with legal experts. You can have a consultation with legal professionals to understand the possible outcomes of web scraping. One of the most important bits of advice is to make sure you're following GDPR(General Data Protection Regulation) or CCPA regulations to avoid legal trouble! Proxies to Future-Proof Your Scraping Proxies to Future-Proof Your Scraping So, what proxies should you choose for web scraping tasks? The best options for reliable, efficient scraping are residential and mobile proxies. By using IPs from genuine devices, residential proxies blend with everyday web traffic, reducing the chances of being flagged by anti-scraping systems. Mobile proxies use IPs from real mobile devices, which are even harder to track and are often ignored by anti-scraping systems. These proxies let you spread traffic across different IPs and locations. Simply using proxies is not enough. We recommend keeping a close eye on the health of your proxies. Regularly monitor their performance for issues like slow response times, blacklisting, or high failure rates. This proactive approach can help avoid disruptions that could hinder your scraping operations. Conclusion Conclusion Web scraping is a valuable tool, but it still comes with challenges due to AI advancements and other factors. Keep on track with all the latest developments, adapt to new rules, and leverage the right tools and strategies. Hope this article will help you handle web scraping hiccups in 2025. Stay tuned with DataImpulse 🚀

Web Scraping in 2025: Staying on Track with New Rules

كشط البيانات من الويب في عام 2025: البقاء على المسار الصحيح مع القواعد الجديدة

Web Scraping το 2025: Μείνετε στο δρόμο με τους νέους κανόνες

El rascado web en 2025: mantenerse al día con nuevas reglas

גירוד אינטרנט בשנת 2025: להישאר במסלול עם כללים חדשים

Web Scraping nan 2025: rete sou wout la ak nouvo règ yo

2025年のウェブスカッピング:新しいルールで追いつく

2025 жылы Web Scraping: New Rules

Pamba la Pamba la 2025: Pamba la Pamba la Pamba la 2025

Веб гребење во 2025 година: Држете се на патот со нови правила

การขูดเว็บในปี 2025: ปฏิบัติตามกฎใหม่

ERROR: invalid target_lang (ts)

Web Scraping vào năm 2025: Giữ theo con đường với các quy tắc mới

Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Methodology 2 Methodology 3 Hardware 3 Hardware 4 Software 4 Software 5 Network 5 Network 6 Consensus 6 Consensus 7 Cryptocurrency Economics 7 Cryptocurrency Economics 8 Client API 8 Client API 9 Governance 9 Governance 10 Geography 10 Geography 11 Case Studies 11 Case Studies 12 Discussion and References 12 Discussion and References A. Decentralization and Policymaking A. Decentralization and Policymaking B. Software Testing B. Software Testing C. Brief Evaluations per Layer C. Brief Evaluations per Layer D. Measuring decentralization D. Measuring decentralization E. Fault Tolerance and Decentralization E. Fault Tolerance and Decentralization 6 Consensus Protocols like Bitcoin [78], Algorand [40], or Ouroboros Praos [78], enable resource holders to engage in the protocol directly with (essentially) whatever amount of resources they have. In these protocols, block producers can, even though they do not have to, form coalitions called pools. In PoW, a pool “leader” validates transactions, and organizes them in a candidate block, while each “member” executes the PoW puzzle for the leader-made block. If a member is successful, the leader collects the block’s reward and distributes it, proportionately to each member’s power. In PoS, the leader has full control over the block’s creation, while the members only pay fees to delegate their staking rights to the leader and collect rewards. Pooling behavior is also driven by temporal discounting [146], i.e., the tendency to disfavor rare or delayed rewards. In essence, a small miner may prefer small frequent payments, at the cost of some fee, over rare large payments, when producing a block. Other systems, like Cosmos [114] and EOS [92], impose restrictions on which parties can participate in consensus and require the rest to delegate their resources to a representative or “validator” node. This “barrier to entry” means that any party without enough stake, i.e., below the system’s threshold or less than its competitors, is required to delegate their staking rights to a validator. At every “epoch”, a committee of (a fixed number of) parties is elected to run the protocol. The election mechanism is voting-based, with resource delegation acting as the voting process. In both types of systems, there are two resources of interest: i) owned participating power, e.g., computational or stake; ii) delegated participating power, including the power to choose a block’s content. Accordingly, the relevant parties are: i) miners and stakeholders, who own hashing power and stake respectively; ii) pool leaders and delegates, who control how the resources are used. Typically, the security of a ledger is guaranteed if the parties that represent an aggregate majority of the participating power are honest (i.e., they follow the protocol as prescribed) [78]. Therefore, the concentration of participating power around few entities poses a threat to the system. This hazard is well-known and blockchain users and participants have actively tried to avoid it since at least 2014 [86]. Those controlling a power majority can hurt liveness, by refusing to publish or accept certain transactions, as well as safety by launching a long-range attack. Both types of attacks also indirectly hurt stability, since the system’s trustworthiness is challenged. A second concern revolves around block proposers. A proposer is a party that maintains a mempool and chooses which transactions are added to a block and in what order. Initially, a single party acted as both block proposer and builder. With the increase in hardware requirements needed to run a full node and the formation of pools, the two roles of proposer and builder were separated. In PoW ledgers, the leader of the pool typically proposes the block’s content, whereas the pool members only run the PoW algorithm. Therefore, pool members are not involved in a block’s construction and often do not even validate its contents. Therefore, the leader may censor transactions (liveness hazard), steal member rewards (stability hazard), or possibly link the user’s resources with information like IP addresses (privacy hazard). In addition, smart contracts enable MEV-type attacks [186], which might hurt stability. Here, block builders have the ability to observe transactions before publication and choose their order in a block, which they can exploit to extract value from honest transactions. A countermeasure that has been introduced is the proposer-builder separation (PBS) model, wherein a trusted party maintains a mempool and proposes a block, whereas validators sign it without ever observing its content (thus not being able to exploit its MEV) [33]. Still, the current implementation of PBS in Ethereum has been criticized for facilitating censorship and centralization, hence its usefulness remains unclear [88]. Finally, a threat arises due to the lack of self-healing, i.e., the inability to recover from a temporary adversarial takeover. In PoW, even if a majority gets corrupted, honest users can increase their own power and, eventually, overthrow the adversary and restore the ledger’s security [10,13]. In PoS though, power shift takes place on the ledger, by transferring stake. If an adversary temporarily obtains a majority, they can prohibit transactions that shift power away from them, thus retaining control indefinitely (for example, a large centralized cryptocurrency exchange can make it hard to issue outgoing payments and withdrawals, while enabling payments between different users of the exchange). Consequently, a diverse stake distribution (cf. Section 7) is vital to protect against takeovers. Authors:
(1) Christina Ovezik, University of Edinburgh (c.ovezik@ed.ac.uk);
(2) Dimitris Karakostas, University of Edinburgh (dkarakos@ed.ac.uk);
(3) Aggelos Kiayias, University of Edinburgh and IOG (akiayias@ed.ac.uk). Authors: Authors: (1) Christina Ovezik, University of Edinburgh (c.ovezik@ed.ac.uk); (2) Dimitris Karakostas, University of Edinburgh (dkarakos@ed.ac.uk); (3) Aggelos Kiayias, University of Edinburgh and IOG (akiayias@ed.ac.uk). This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv available on arxiv

Can Blockchain Stay Fair for the Long Haul?

AI agents and bots increasingly dominate digital spaces, and distinguishing real human interactions from automated systems has become a pressing challenge. Steven Smith, Head of Protocol and Applied Research at Tools for Humanity, is in charge of solving this through the World project. With the recent mainnet launch of World Chain—a blockchain designed to prioritize verified humans—Steven and his team are pioneering proof of human technology to ensure fairness and trust in digital ecosystems. In this interview, we explore the technical breakthroughs behind Priority Blockspace for Humans (PBH), the role of World ID in combating AI-driven challenges, and the future of trust-first AI ecosystems that empower human agency. 1. Ishan Pandey: Steven, as the Head of Protocol and Applied Research at Tools for Humanity, you’re driving the World project’s mission to prioritize real humans in a digital world overrun by AI agents and bots. Can you tell us about the vision behind World Chain? Steven Smith: World Chain, our blockchain designed for real humans, was built for a few core reasons. First, World Network has grown substantially, and we anticipate even greater scale in the future – as such, we need to develop the best infrastructure to ensure that scaling is seamless. Second, we envision a future where humans can coexist with AI agents and bots. Contrary to the common perception that bots are inherently bad actors, many are essential to network operations. AI agents, in particular, serve as productivity enhancers and multipliers, amplifying what individuals can achieve. Steven Smith: The key is ensuring that humans remain central to any given network. World Chain enables this by creating a blockchain specifically designed to accommodate both AI agents and bots while insulating human users from their potential drawbacks. It’s a blockchain built for people – a human-centric network that prioritizes verified users. Since World Chain’s mainnet launch in October 2024, 24+ million World ID holders across 160+ countries have either migrated or are in the process of migrating to World Chain, making it the largest blockchain of individually verified humans in the world. World Chain is also maintaining the highest UOPS/TPS ratio of any Ethereum blockchain according to L2Beat. The network is supported by world-class blockchain service providers like Optimism, Alchemy, Uniswap, Safe, Dune, and Etherscan, and it integrates World ID natively to prioritize human transactions. 2. Ishan Pandey: What were some of the biggest technical obstacles you faced when designing Priority Blockspace for Humans (PBH) to ensure that verified human transactions take precedence over those from bots or AI agents? Steven Smith: The fundamental challenge in designing Priority Blockspace for Humans (PBH) was determining, at the blockchain consensus level, whether a transaction originates from a human. Fortunately, World ID, which is a proof of human credential used onchain, solves this problem. World ID leverages zero-knowledge proofs (ZKP), allowing users to prove their unique humanness without revealing personal data. PBH, which recently launched on the World Chain testnet, is like a dedicated highway lane for humans, ensuring smooth and prioritized transactions. Every human-verified transaction on World Chain carries an associated ZKP, enabling the network to identify and prioritize them. Steven Smith: Another consideration was how to architect PBH as a modular, generic service that could extend beyond its initial application. Our collaboration with Flashbots and consultation with Optimism and Paradigm played a crucial role in refining this approach. By cleanly separating the block-building mechanism from the execution client, we designed an architecture that can be adapted for various use cases beyond PBH. 3. Ishan Pandey: Can you walk us through how PBH works within World Chain’s architecture? How does it leverage World ID and the OP Stack to prioritize human transactions while maintaining security and decentralization? Steven Smith: PBH is currently live on testnet and is built on OP Stack, the rollup SDK from Optimism. OP Stack consists of multiple infrastructure components, one of which is responsible for block construction, deciding which transactions get included. Steven Smith: To implement PBH, our collaborations integrated the Reth execution client to modify the block-building process. PBH transactions carry a special payload that includes a World ID proof, ensuring that a transaction comes from a verified human. This mechanism enables priority inclusion of human transactions while maintaining decentralization and security. 4. Ishan Pandey: With World Chain now live on mainnet, how does your system handle a high volume of transactions while ensuring that verified humans aren’t outpaced by automated systems? What optimizations were key to achieving this balance? Steven Smith: The core optimization is reserving a portion of every block specifically for human transactions. These transactions are prioritized at the top of each block, ensuring timely processing and insulation from bidding wars. Automated systems, particularly bots, operate with economic incentives, often paying high fees to secure priority inclusion. Without intervention, human users would face rising costs and delays. Steven Smith: PBH eliminates this problem by guaranteeing a set amount of blockspace for human-verified transactions, ensuring they remain affordable and accessible regardless of network congestion. Additionally, verified users benefit from a gas allowance, helping to cover the cost of casual transactions and ensuring affordability for everyday users. Casual transactions make up a majority of activity on World Chain, and this is what makes the network unique. We are bringing all people onchain, not just power users. 5. Ishan Pandey: PBH and World ID have the potential to act as a global “blue check mark” for AI-driven services. How do these technologies enable safe and fair integration with other blockchains or platforms, and what impact do you foresee on the broader ecosystem? Steven Smith: PBH was designed with a specific application in mind – ensuring a special class of service for humans on World Chain. However, the underlying architecture is highly modular and can be adapted by other chains and platforms. The PBH code is open-source, allowing any blockchain to implement similar human-prioritized transaction mechanisms. Steven Smith: World ID functions as a global “blue checkmark” for unique humanness, providing a way to distinguish real users from automated entities across AI-driven services. In addition, it empowers users by allowing AI agents acting on their behalf to inherit the same verified privileges. The open-source nature of World ID ensures it can propagate across other chains and applications, further solidifying its role in civil resistance and human verification. Examples of World ID integrations include apps and services like Minecraft, Reddit, Telegram, Shopify, and Mercado Libre, as well as blockchain projects like DRiP, DSCVR, Galxe, and more. Most recently, World announced a global partnership with gaming company Razer to launch ‘Razer ID Verified by World ID,’ a secure, single sign-on proof of human technology designed to create a safer and more authentic gaming environment. These integrations demonstrate World ID’s growing role as a trust layer across Web3, AI-driven platforms, and beyond. 6. Ishan Pandey: Looking ahead, what technical advancements are next for the World project? How do you plan to further enhance human agency and tackle challenges like deepfakes, scams, and agent swarms in trust-first AI ecosystems? Steven Smith: We are continuing to develop innovative mechanisms to enhance human agency onchain. One major initiative is expanding the gas allowance for verified human users to cover a portion of their onchain activities, minimizing the friction of interacting with blockchain applications. We are also exploring new fee market structures to further insulate verified human transactions from volatility. Steven Smith: Beyond blockchain transactions, we’re actively working on technologies to combat deepfakes, scams, and agent swarms. This includes expanding World ID Credentials and developing World ID Deep Face, a tool designed to combat deepfakes in real-time video and chat interactions. Ultimately, our goal is to build a blockchain ecosystem that prioritizes human agency, ensuring that individuals remain at the center of digital trust and economic participation. Don’t forget to like and share the story! Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYO Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYO Vested Interest Disclosure: business blogging program business blogging program

How World Chain Ensures Humans Stay in Control: The Blockchain Built for Verified Humans

كيف تضمن سلسلة العالم بقاء البشر في السيطرة: بلوكتشين مصممة للبشر المعتمدين

Як World Chain гарантуе, што людзі застаюцца пад кантролем: блокчейн, створаны для правераных людзей

Jak světový řetězec zajišťuje, aby lidé zůstali pod kontrolou: Blockchain vytvořený pro ověřené lidi

Cómo World Chain garantiza que los humanos mantengan el control: la blockchain diseñada para humanos verificados

Como a cadea mundial garante que os humanos manteñan o control: a cadea de bloques creada para humanos verificados

वर्ल्ड चेन कैसे सुनिश्चित करता है कि मनुष्य नियंत्रण में रहें: सत्यापित मनुष्यों के लिए निर्मित ब्लॉकचेन

World Chain が人間による制御を確保する方法: 認証された人間向けに構築されたブロックチェーン

World Chain이 인간이 통제권을 유지할 수 있도록 보장하는 방법: 검증된 인간을 위해 구축된 블록체인

Как World Chain обеспечивает людям контроль: блокчейн, созданный для проверенных людей

Hur World Chain säkerställer att människor håller kontrollen: Blockchain byggd för verifierade människor

உலகச் சங்கிலி மனிதர்கள் கட்டுப்பாட்டில் இருப்பதை எவ்வாறு உறுதி செய்கிறது: சரிபார்க்கப்பட்ட மனிதர்களுக்காக உருவாக்கப்பட்ட பிளாக்செயின்

Чӣ гуна занҷири ҷаҳонӣ одамонро дар назорат нигоҳ медорад: Блокчейн барои одамони тасдиқшуда сохта шудааст

Stories, News & Mentions about STAYY

Categories

Trending Topics

Stories, News & Mentions about STAYY

Stayy’s Stories on HackerNoon

Categories

Trending Topics