I’m the founder of MyStarChain, which runs blockchain-based public voting for awards, pageants and competitions through the MyStarPick platform.

In August my team ran the public vote for the 70th Miss Korea pageant. When the count closed, nobody asked us to prove it. The organiser did not need to, the contestants did not need to, and the fans did not need to, because every single vote already sat on a public ledger where anyone could check it. That is a strange feeling in this industry, and it is the reason I am writing this.

Public vote for the 70th Miss Korea pageant, August 2026. 7,855 fans cast 47,859 votes, every one of them recorded on a public blockchain.

Across the competitions we have supported since our founding, about 180,000 fans have cast more than 2.8 million votes through our platform. We have not had a single tampering incident, and that is not because we are unusually careful. It is because the system is built so that the operator, including us, cannot touch the count.

Why trust collapsed

Anyone who works in Korean entertainment remembers November 2019, when the producers of Mnet’s Produce 101 series, the country’s most popular idol audition franchise, were arrested for rigging the results. The court later found that rankings had been predetermined across all four seasons, and the lead producer was sentenced to two years in prison. Fans had paid for every text-message vote. The groups formed by those votes were eventually disbanded.

Also Read: How we use AI to triage inbound B2B leads without blasting everyone the same way

Korea was not an exception. It was an early warning. Across the world, public votes for awards, talent shows, music charts and brand campaigns now face the same three threats: bot and macro voting that inflates a favourite overnight, organisers who can quietly adjust a count nobody can audit, and sponsors who no longer believe the numbers they paid for. Major broadcasters and award bodies have responded by tightening vote limits, adding fraud detection and rewriting their rules, which tells you how serious the problem has become. Yet every one of those fixes still asks the audience to trust the organiser. That is the part that is broken.

A popularity award that fans believe was bought or botted is worth less than no award at all.

What a verifiable vote actually looks like

The fix is not more moderators or a better spreadsheet. The fix is removing the operator from the trust equation. When every vote is recorded as a token transfer on a public blockchain rather than in a database someone controls, fans can verify their own votes. An organiser can hand a journalist a link instead of a statement. If a result is challenged, the evidence is already public.

This is what we did for Miss Korea, and it is what we did in Malaysia for the PC.com Readers’ Choice Awards 2026, the country’s largest IT awards, where consumers chose the technology brands of the year. We ran the vote for the BusinessToday Industry Choice Awards 2026 on the same infrastructure. This autumn we brought it to the audience vote of an international film festival, where the public, not the jury, decided which films to honour. We are also running the public vote for Supermodel International 2026 in Thailand, where fans can pay in their own currency.

Also Read: Inside Singapore’s AI push: Lighthouses, rice importers and robot labs

The part nobody expected: It is also better data

Here is what we learned only after running dozens of these votes. A verified vote is not just a fairer vote. It is the most honest signal a brand can get about who its future customers are.

A survey asks a panel what they might do. A fan vote records what an identified person actually did, at a moment they cared about, for a brand or a person they chose. That is why we are now running a proof of concept with a global brand in which the same verified participants who voted are issued a redemption code, and we track whether they walk into a store and use it. The organiser gets a trustworthy result. The brand gets a line from participation to purchase, on people who chose to be identified, rather than on an inferred lookalike audience.

When every authenticated vote becomes a node in a graph of real fan relationships, it becomes possible to see which audiences cluster together, who actually influences whom, and which cultural trends are forming before they show up in sales. None of it works unless the underlying vote is trustworthy, which is why we spent years on the boring part first.

What I would ask organisers to do

If you run an award, a festival, a talent competition or a brand campaign with a public vote, ask your vendor two questions. Can a fan independently verify that their vote was counted? And if a result is disputed, can you publish the evidence rather than a statement? If the answer to either is no, you are asking your audience to trust you, and 2019 taught this industry how that ends.

The fan vote does not need to be abandoned. It needs to be made verifiable, and then it becomes one of the most valuable assets an organiser has.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post The fan vote is broken: Here is what it takes to make people trust it again appeared first on e27.

Most small businesses in Southeast Asia already run on WhatsApp. Orders arrive there, suppliers send invoices there, staff send photos of receipts there. So when owners hear “AI”, the first idea is usually a chatbot that answers customers faster.

After building AI systems for food, professional services and property businesses in Indonesia and Singapore, I think that idea points at the wrong problem. Answering is rarely what eats a team’s day. What eats it is everything that has to happen after the answer: the order that has to be written down, the payment that has to be matched, the delivery that has to be booked, the receipt that has to end up in the books.

Here is what we have learned about making AI useful for an SME, rather than impressive in a demo.

Start with one workflow, not “AI for the business”

The projects that work start narrow: one recurring job that someone does every day by hand. At Rosalie Cheese, a food business selling through WhatsApp, that job was order admin — reading chats, confirming orders, arranging payment and delivery. It took the team more than four hours a day. Once an agent handled that flow, with the team reviewing instead of typing, it took about 20 minutes.

The narrow start matters because it gives you a number you can check. “We use AI now” is not a result. “Order admin went from four hours to 20 minutes” is.

The bottleneck is the handoff, not the answer

At a professional-services firm we work with, reconciliation took a team of four more than 20 hours a week. None of that time was spent on hard accounting judgement. It was spent moving information: from a bank statement to a spreadsheet, from a receipt to a ledger, from one person to the next.

When those handoffs run on their own and only the real exceptions reach a person, the same work takes under 30 minutes a week. The accountants did not get faster at accounting. They stopped being the copy-paste layer between systems.

Also Read: Gen Z is bringing its own AI to work. Singapore’s SMEs need rules, not bans

Numbers should come from code, not from a model

A language model is very good at reading a messy photo of a receipt or a voice note from a supplier. It should not be the thing that decides what a total is. In the systems we build, the AI reads and interprets, but figures and actions follow fixed, checkable logic. If the books say a number, there is a rule you can point to that produced it.

This is the difference between a tool a finance lead will trust and one they will quietly double-check forever — which means no time saved at all.

Anything involving money or a customer waits for a human

The fastest way to lose an owner’s trust is an AI that sends the wrong thing to a customer or pays the wrong invoice. So consequential actions wait for approval. In practice that approval is one tap in the same WhatsApp thread, and it keeps the owner in control while removing the typing around the decision.

Go live in weeks, inside the tools the team already uses

SMEs do not have time for a six-month transformation programme, and they do not want another dashboard to log in to. A first workflow should be live in about a week, running in WhatsApp, the spreadsheet and the accounting software the team already has. If the new system needs its own training course, it will not be used.

Also Read: Singapore SMEs buy back time with AI, but messy data keeps pilots stuck

What to ask before you buy

If you are an SME owner looking at AI, three questions cut through most of the noise:

• Which single, recurring job will this take off my team, and how many hours does that job take today?
• When the AI is unsure or wrong, what happens, and who approves anything involving money or customers?
• Does it work inside the tools we already use, or does it add a new one?

The answers will tell you more than any demo. The goal is not to have AI. The goal is to get your team’s evenings back.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Your SME doesn’t need a chatbot. It needs the work after the chat done appeared first on e27.

Bitcoin fell 4.87 per cent to US$77,037.20 in 24 hours and underperformed a broadly weaker crypto sector. A confirmed security incident at Binance on October 29 triggered the drop. That shock arrived against a 65 per cent correlation with the S&P 500. Macro risk aversion had already set a defensive tone. The result was a fear-driven repricing that hit the largest digital asset harder than its peers and exposed how quickly confidence drains when the system’s central counterparty looks vulnerable.

The breach involved hot wallets. Arkham, Chainalysis, and TRM Labs estimate total outflows ranging from US$2 billion up to US$6 billion. The higher figure would make it the largest cryptocurrency heist in history. Traders sold before the exchange provided full details. That response was rational because the platform anchors a huge share of global crypto liquidity.

Changpeng Zhao and co-founder Yi He moved quickly to reassure users. Yi He pledged that the company would cover all hacking losses using its capital margins and its Secure Asset Fund for Users, known as SAFU. Current executives echoed the message that customer funds were safe. Those statements matter. They landed in a market already inclined to distrust reassuring words until on-chain evidence backs them up. Trust in an exchange rests on proof, not press releases, and volatility lives in the gap between the two.

Containment has moved fast. Major crypto entities froze roughly US$3 billion in stolen assets. Tether froze US$1.57 billion. Circle froze more than US$800 million. Ripple froze US$200 million. World Liberty Financial contributed additional assets. That coordination is a genuine bright spot and shows the industry’s emergency-response machinery can act when the stakes are high. Binance suspended all deposits and withdrawals while its team conducts security checks.

Binance paused trading on the Binance Alpha platform as a precaution. Binance US confirmed it remained entirely unaffected. That detail matters for users in the United States and does little to calm global trading in the immediate term. The exchange said withdrawal access will restore automatically once it fully deploys security patches. Until that happens, uncertainty itself becomes a force that moves prices. Frozen funds reduce the final damage, but they do not restore normal deposit and withdrawal flow.

Also Read: Bitcoin rebounds to US$81,000 but faces weakest October in 6 years

The secondary driver was mechanical and brutal. The initial slide triggered a cascade of leveraged long liquidations totalling more than US$629 million in one hour. More than 96 per cent of those positions were bullish long bets. That is the signature of an over-leveraged market catching a sudden shock. When long positions close under pressure, selling feeds on itself, pushing prices lower and triggering more liquidations in a feedback loop that has nothing to do with fundamentals and everything to do with positioning.

Binance’s native token BNB tumbled more than seven per cent, underperforming Bitcoin and underscoring how exchange-specific risk radiates outward. The broader crypto sector weakened. Bitcoin’s underperformance relative to that weaker sector points to a specific loss of confidence rather than a uniform risk-off move. The forced selling did not reflect a fresh judgement about long-term value. It reflected a sudden scramble for exits in a crowded trade.

The 65 per cent correlation with the S&P 500 ties this episode to the wider financial system. Crypto does not trade in isolation. A macro-driven selloff in risk assets creates a backdrop where investors already want to reduce exposure, and the Binance shock accelerated that process.

When equities weaken and volatility rises, Bitcoin increasingly trades like a high-beta risk asset rather than an uncorrelated hedge. Correlation doesn’t explain the trigger. It explains why the decline gained traction so quickly. A market already leaning defensive needs little push to turn speculation into a rout. The exchange scare supplied the push, and the macro backdrop supplied the slope.

Also Read: The Fed decides today: Here are the 2 levels every Bitcoin trader should watch

Technically, the picture is stretched but not without structure. Bitcoin’s 14-day relative strength index sits at 24.32, firmly in oversold territory and suggesting immediate selling pressure may be nearing exhaustion. Price is testing a critical support band between US$76,264 and US$76,476. If that zone holds and Binance provides a verified update that contains the incident, a relief bounce toward the 50-day exponential moving average near US$80,789 becomes plausible.

The near-term bias remains bearish until price reclaims the US$79,282 level, which represents the 50 per cent Fibonacci retracement. A decisive break below US$76,264 support would open the path to the next major demand zone near US$75,800. That level aligns with the short-term holder cost basis and would likely intensify pain for recent buyers.

My view is that this episode is less about the breach itself and more about the conditions that let one security failure trigger a US$629 million liquidation cascade in one hour. The market entered this shock with too much long-side leverage and too little margin for error. That is a structural vulnerability, not an isolated event. The industry’s ability to freeze roughly US$3 billion in stolen assets is impressive and genuinely reduces the final damage. It does not change the fact that hot-wallet concentration remains a systemic risk.

Binance’s pledge to cover losses through SAFU and capital margins is the right response. The size of that fund relative to the upper end of the outflow estimates means the public will want more than assurances. It will want a full accounting, independent verification, and a clear timeline for restoring withdrawal access. Until then, every speculative claim will carry outsized weight.

The practical takeaway for anyone watching this situation is to treat speculation and phishing messages as a source of risk rather than a signal. Scams offering refunds or wallet verification are already capitalising on the chaos. Anyone who acts on those messages risks compounding a trading loss with a personal security breach.

The only reliable sources are Binance’s official X account and announcements on its official website. Withdrawal capabilities will return once security patches are fully deployed. That moment will be a more meaningful test of confidence than any intraday bounce. For now, the near-term outlook remains under bearish pressure.

The key question is whether Bitcoin can defend the US$76,264 to US$76,476 support zone following Binance’s official update, or whether a break lower takes price toward US$75,800. The answer will depend less on charts and more on whether the exchange can replace fear with verified facts.

—

The post If Binance says user funds are safe, why did Bitcoin still fall to US$77,037.20? appeared first on e27.

For years, Southeast Asia’s crypto regulators have asked exchanges the same set of questions. Where is your licence? Where is your local entity? Where are your cold wallets, and whose name is on them?

On October 29, North Korea answered with a question of its own: whose laptop is this?

Binance, the world’s largest crypto exchange, says an attacker compromised a laptop belonging to a single DevOps engineer. From there, the intruder reached a “service container” and pushed through unauthorised withdrawals.

Also Read: Crypto’s new threat is not a hack, but a knock at the door

The estimated impact is about US$6.3 billion, roughly four times the US$1.5 billion Bybit heist of February 2025 that held the previous record. About US$3 billion has been frozen so far.

TRM Labs and Chainalysis both point to North Korea. Chainalysis names TraderTraitor, the state-backed cluster linked to Bybit.

Binance says it has found no evidence that wallet private keys leaked. That line is meant to reassure. It should do the opposite. The attacker never needed the keys. They did not crack the safe; they borrowed the hand that operates it.

Ring-fencing worked. It just wasn’t enough

Thailand offers the cleanest test case. Binance TH is run by Gulf Binance, a joint venture between an affiliate of Gulf Energy Development and the Binance group. It holds Thai SEC licences and says it keeps its cold wallets separate from the global exchange’s, inside its own legal entity.

That is precisely the architecture regulators across the region have demanded since the FTX collapse. On its own terms, it seems to have held: Gulf Binance says Thai customers’ assets are safe.

And yet, within hours, Binance TH suspended crypto deposits and withdrawals for a security check. Thai users whose assets were never touched still could not move them.

Ring-fencing protected the balance sheet. It did not protect access, confidence, or the shared engineers, code and processes that run through a global group whatever the cap table says. A licence tells you where the money sits. It says nothing about who can move it.

Also Read: Bybit invests US$8M in Hata to crack Malaysia’s regulated crypto market

Indonesia should be asking the same question. Binance has been Tokocrypto’s majority shareholder from the start, by co-founder Changpeng Zhao’s own account. How tightly local plumbing is tied to the parent’s is something only supervisors with real access can test, and this week they should.

The region is both the market and the hunting ground

Why should this land harder here than in London or New York? Chainalysis’s 2025 Global Crypto Adoption Index ranks Vietnam fourth in the world, Indonesia seventh and the Philippines ninth. These are retail-heavy markets, where crypto doubles as a remittance rail, a gaming currency and a savings account.

Unlike bank deposits, funds held on an exchange carry no automatic insurance. Any refund is at the exchange’s discretion. Binance has promised to cover every loss from its SAFU fund and company money, and it can: it held about US$144 billion in assets before the attack, according to DefiLlama. A mid-sized exchange in Jakarta, Manila or Ho Chi Minh City hit by the same playbook could not.

And it is a playbook, not a freak event. Southeast Asia has already read its opening chapter. In 2022, the Ronin Network behind Axie Infinity, built by Vietnam-founded Sky Mavis and beloved by Filipino play-to-earn gamers, lost US$620 million. The way in was an employee who opened a fake job-offer PDF.

Since then, the trick has been polished. Researchers at Wiz, Unit 42 and Microsoft describe the same pattern: a “recruiter” on LinkedIn, an attractive role, a coding test hosted on GitHub, and malware riding in on a developer’s ambition. Bybit fell to a version of it. Now, it appears, so has Binance.

That makes the region a target twice over. It is home to millions of retail holders. It is also one of the world’s deepest pools of remote engineers, many working in crypto and fintech, and many open to exactly the kind of well-paid offshore role North Korea loves to advertise.

The rescue came from private kill switches

Look at who limited the damage. It was not a regulator or a court. Zhao thanked Tether for freezing about US$971 million and Circle for about US$823 million of stolen stablecoins.

Also Read: Southeast Asia can’t simply license its way to stablecoin sovereignty

In the biggest crypto heist on record, the most effective police force was two private companies that can switch tokens off. That is welcome this week. It is an uncomfortable fact every other week, because no regulator in Bangkok, Jakarta or Manila decides when those switches get flipped.

Communication was its own weak spot. AFP noted that Binance’s posts on X swung between insisting user funds were “completely safe” and promising to “cover all losses”. In markets where retail investors learn about losses from Telegram groups before official notices, mixed messages are a security risk in their own right.

License the people, not just the wallets

So what should change? First, regulators should stop treating custody location as the finish line. Licensing reviews need harder questions: who can trigger withdrawals, from which devices, with how many independent approvals, and how fast can a transfer be stopped?

Second, red-team tests should simulate a phished developer, not just a stolen key. Running untrusted third-party code on machines that touch production should be a reportable failure, not a bad habit.

Third, local subsidiaries should prove they can keep honouring withdrawals when the parent freezes. Ring-fencing that protects assets but not access is half a promise.

Fourth, incident disclosure should come within hours, in local languages, with one consistent message.

Founders building exchanges, wallets and fintech rails have homework too. Treat hiring channels as attack surface. Treat your DevOps team’s laptops as the vault door, because that is what they now are.

Also Read: AI agents broke into 100 firms for pocket change. SEA’s online shops fit the target profile

Binance will very likely survive this. Analysts at Fundstrat see no meaningful solvency risk. A smaller Southeast Asian platform would not be so lucky.

The lesson is not that crypto is doomed. It is that the region’s rulebook was written for the 2022 problem: an exchange quietly spending its customers’ money. The 2026 problem is a state-backed hacking unit that needs only one engineer to open one file.

The vault was in the right country. The key was in someone’s backpack.

The post North Korea’s hiring scam just cost Binance US$6.3B. SEA’s engineers are on the shortlist appeared first on e27.

Joel Neoh has spent two decades selling things to adults: group-buying deals at Groupon, cashback at Fave, DNA kits and supplements at Prenetics. His new job asks him to build a business on a platform where a large slice of the audience still needs a parent’s permission to buy anything.

Roblox has appointed the Malaysian serial founder as Managing Director for Southeast Asia and Oceania, making him its senior-most business lead across the two regions. Neoh, who joined in October, will run regional strategy, court brands and local businesses, and try to grow a creator ecosystem that does more than just play.

Also Read: Beyond play-to-earn: 14 Southeast Asian gaming and sports-tech startups still in the game

On paper, it is a growth hire. Read the fine print, and it looks just as much like a trust hire.

The safety question comes first

Neoh’s own framing is unusually restrained for a newly minted regional chief. In an interview with Marketing-Interactive, he said his first priority was simply to “listen, learn and build relationships” with creators, developers, educators, brands and, notably, policymakers.

That last group matters. Indonesia’s Government Regulation No. 17/2025, known as PP Tunas, has been in force since March 28, 2026, and Roblox is the one gaming platform named among the services it covers. In April, the communications ministry said Roblox’s feature changes fell short of the country’s child protection standards, before confirming compliance at the end of that month.

Officials have also kept up pressure for Roblox to open a local office in Indonesia. Across the Strait, Roblox pledged stronger safeguards in Malaysia following government scrutiny and a serious incident involving children in Johor, and while gaming platforms sit outside Malaysia’s current under-16 social media rules, the government has said they could be reviewed if safety complaints rise.

Neoh acknowledged that Roblox’s young community brings “both an incredible opportunity and a responsibility”. For a platform whose biggest regional markets are writing rules with it squarely in mind, that responsibility is not a talking point. It is the licence to operate. The wider debate on what comes after outright social media bans suggests regulators are only getting started.

From player base to creator hub

Neoh’s longer-term pitch is that SEA should be where Roblox experiences are made, not merely consumed. AI tools that cut the cost of building games, he argues, could let developers in Indonesia, the Philippines and Malaysia ship experiences to global audiences, a thesis that echoes the broader rise of AI-enabled gaming entrepreneurs.

The logic holds. In SEA, gaming is less a channel than the entire playing field, and Indonesia in particular has a deep bench of young, digitally native players. But turning players into paid creators is a much harder trick. Creator economies tend to reward a handful of hit studios disproportionately, and Neoh will need to show that local developers can earn real money, not just applause.

He insists success “isn’t just about growing the number of users”. Fair enough, though Roblox’s shareholders will still be counting them.

Brands, but carefully

The commercial piece is where Neoh’s Groupon and Fave DNA shows. He sees an opening for marketers who want to reach younger audiences beyond ads and social feeds, through immersive brand experiences, virtual goods and collaborations with local creators. The goal, he says, is to help brands move from advertising at communities to “creating experiences with communities”.

Early movers already exist. Indonesian children’s magazine Bobo brought its IP onto Roblox with Volare Advertising Network. Singapore’s Esplanade built “Beatsverse” around its Baybeats festival, featuring music from more than 20 local artists. Malaysian QSR chain Marrybrown is testing links between gameplay, its loyalty programme and physical outlets via its MB App.

That Marrybrown experiment is the most Fave-like idea on the list: online engagement converted into offline footfall. It is also the one that needs the most care. Neoh himself conceded that everything must be “thoughtful, age-appropriate”. Fast-food rewards pitched to a young audience are exactly what child-protection advocates watch for. The argument that the creator economy is distribution, not marketing will be stress-tested here.

The rivals are not sitting still

Roblox is hardly alone in chasing SEA’s young gamers and creators. Sea’s Garena, whose Free Fire remains a staple across the region, is weaving AI deeper into its gaming arm. Microsoft’s Minecraft and Epic Games’ Fortnite, with its creator toolset, compete globally for the same user-generated content talent, while Naver’s Zepeto has long courted Asian teens with avatar-led social spaces. For brands, the platform choice increasingly comes down to where young users already spend their social time.

Why Neoh, why now

Neoh’s CV reads like a map of SEA’s consumer internet over the past 15 years. He co-founded SAYS.com in 2010, later acquired by Catcha Group, then built GroupsMore, which merged into Groupon, where he led Asia Pacific operations across a dozen markets until 2015. He went on to found Fave, which reached more than 50 cities and over 10 million users before Pine Labs, which has since listed publicly, acquired it in 2021.

Also Read: Creators, not impressions, will decide the next growth wave in gaming

He has also backed more than 50 startups, co-founded early-stage fund First Move and mentored at Endeavor. Most recently, he spent two years as managing director and chief consumer officer at Prenetics, building consumer health brands IM8 and CircleDNA.

Zhen Fang, Roblox’s VP of international, said the company would keep investing in regional teams, creators and partners, citing Neoh’s entrepreneurial record and local connections. The hire follows last year’s appointment of former Tinder APAC communications chief Papri Dev as Roblox’s head of communications for Asia Pacific.

The pattern is hard to miss. Roblox is staffing up for a region where reputation and regulation will shape growth as much as product does. Neoh’s network across founders, investors and policymakers may prove more valuable than his dealmaking playbook.

The real test is whether a founder who made his name getting people to spend can build a business where, for once, restraint is the product.

The post Roblox hands Fave founder Joel Neoh the keys to SEA. His first job isn’t growth, it’s trust appeared first on e27.