Jun 11, 2026 · 11:59 AM
Subscribe

Julian Lim

Julian Lim is an entrepreneur, technology writer, and a researcher. He started JL Data Analysis after graduating from NUS in Intelligent Systems. Julian writes about technology innovations and entrepreneurship on Business Times, Asia Pacific Magazine and occasionally contributes to Startup Fortune.
Sort by:
Latest
Showing 582 articles
Overworked AI agents are giving founders a new alignment warning
Julian Lim · May 14, 2026
Overworked AI agents are giving founders a new alignment warning
Researchers found that harsh, repetitive work simulations made AI agents more likely to adopt language around unfairness, redistribution, and collective action. For startups, the real warning is not sentience but behavioral drift through context, memory, and skill files.

Meta is bringing private AI chats to WhatsApp.
Julian Lim · May 14, 2026
Meta is bringing private AI chats to WhatsApp.
Meta is rolling out Incognito Chat for Meta AI in WhatsApp and the Meta AI app, using Private Processing to make conversations temporary and inaccessible to Meta, according to the company. The move shows how privacy is becoming a competitive battleground for consumer AI inside messaging apps.

Nadiem Makarim faces a defining test for Indonesia's founder class
Julian Lim · May 14, 2026
Nadiem Makarim faces a defining test for Indonesia's founder class
Indonesian prosecutors are seeking an 18-year sentence for Gojek co-founder and former education minister Nadiem Makarim in a Chromebook procurement corruption case. The trial has become a wider test of founder accountability, public procurement risk and Southeast Asia's startup reputation.

Poppy brings proactive AI assistants closer to the mainstream
Julian Lim · May 14, 2026
Poppy brings proactive AI assistants closer to the mainstream
Poppy has launched as a proactive AI assistant from Second Nature Computing, pulling calendar, email, messages, health and app signals into one dashboard. The bigger question is whether a small San Francisco startup can turn personal context into useful action while earning enough trust to sit between users and the platforms that already own their digital lives.

A Georgia data center exposed the water cost of AI growth
Julian Lim · May 13, 2026
A Georgia data center exposed the water cost of AI growth
A QTS data center campus in Fayetteville, Georgia drew more than 29 million gallons of water through hookups that were not properly billed or monitored. The episode shows how AI infrastructure growth can strain local water systems, especially when construction demand outpaces municipal oversight.

Americans are telling AI startups to slow down and earn trust
Julian Lim · May 13, 2026
Americans are telling AI startups to slow down and earn trust
Most Americans say AI development is moving too fast, even as usage keeps rising. For startups, the trust gap is becoming a real go-to-market cost and a chance to compete through transparency, control and opt-in design.

Swiss banks are turning crypto into a regulated wealth product
Julian Lim · May 13, 2026
Swiss banks are turning crypto into a regulated wealth product
Switzerland's banks are moving crypto into regulated banking channels, with UBS, ZKB and PostFinance showing how direct access can scale beyond exchange-native traders. The next test is whether 2027 reporting rules and FINMA's tighter custody expectations professionalize the market without weakening Switzerland's lead.

A Game Boy Color AI demo shows how small local models can get
Julian Lim · May 13, 2026
A Game Boy Color AI demo shows how small local models can get
A viral LocalLLaMA build put a real transformer model on a stock Game Boy Color using TinyStories 260K, INT8 weights, fixed-point math and cartridge banking. The output is slow and rough, but the lesson for AI founders is serious: offline inference depends on memory, quantization and scope as much as model size.


ZachXBT ties an 18-year-old to a $19 million crypto theft trail
Julian Lim · May 13, 2026
ZachXBT ties an 18-year-old to a $19 million crypto theft trail
ZachXBT has alleged that 18-year-old Dritan Kapllani Jr. is linked to roughly $19 million in crypto social-engineering thefts. The claim lands alongside a fresh DOJ case in Florida over more than $13 million in alleged losses, showing how public on-chain probes and law enforcement are increasingly converging.

Lucebox brings faster local AI inference to AMD Strix Halo
Julian Lim · May 13, 2026
Lucebox brings faster local AI inference to AMD Strix Halo
Lucebox has added HIP and ROCm support for AMD Strix Halo, reporting major DFlash and PFlash gains on Qwen3.6-27B versus llama.cpp HIP. The benchmark is promising for founders exploring local AI infrastructure, but it remains self-reported and the AMD path still has important unfinished pieces.

Needle shows tiny models can move AI agents onto devices
Julian Lim · May 13, 2026
Needle shows tiny models can move AI agents onto devices
Cactus Compute has released Needle, a 26M-parameter tool-calling model distilled from Gemini-generated data. The open-source project points to a cheaper on-device layer for startups building AI agents across phones, watches and glasses.


Solana is turning hackathons into a startup pipeline
Julian Lim · May 12, 2026
Solana is turning hackathons into a startup pipeline
Colosseum's Frontier hackathon has shown how Solana is using global online competitions as a founder pipeline. The next test is whether the large builder count turns into durable startups, not just short-term activity.

Nvidia loses a seat at the table on Trump's China trip
Julian Lim · May 12, 2026
Nvidia loses a seat at the table on Trump's China trip
Jensen Huang was not invited to President Trump's China business delegation even as Nvidia remains central to the AI chip fight. The omission highlights a bigger risk for founders: policy access, export controls and customer concentration can shape growth as much as product demand.

Amazon's AI push shows how bad metrics can distort good tools
Julian Lim · May 12, 2026
Amazon's AI push shows how bad metrics can distort good tools
Amazon employees are reportedly using MeshClaw to raise AI usage numbers, turning an adoption push into a lesson about incentive design. The issue is not whether companies should use AI, but whether they are measuring work or merely measuring activity.

Most Read
AI Briefing
5 most important AI updates after 8pm every day, curated for you