Singapore AI news archive.
Every SG + AI news item we classified, grouped by ISO week. Each card carries the analyst one-liner, category tag, and sentiment dot that appeared on the live reel. Kept permanently so we can look back on how the ecosystem evolved.
About this archive: each item is ingested and classified automatically every day with a category (policy, enterprise, funding, talent, infrastructure or products), a sentiment read, and a one-line analyst note. The archive is kept permanently. The same feed powers the news sections on the altronis.sg homepage and the @AIinSG Telegram channel.
Week 33 · 2026
33 items
ProductsWIR · AI Is Helping Solve the Intricate Genetic Puzzle of Schizophrenia
AI models mapped 766 schizophrenia genes, shifting research from single mutations to network-level biological interactions.
ProductsSCMP · Meta to challenge China’s open-weight AI dominance amid US regulatory fears
Meta positions open-weight models as a geopolitical hedge, shifting the competitive axis from pure performance to compliance risk for US enterprises.
ProductsAIN · How AI is changing the vulnerability response timeline
AI-assisted exploit generation forces a shift from signature detection to behavioral analysis in vulnerability response.
ProductsQCR · Podcast with Julien Camirand Lemyre, Chief Executive Officer & Co-founder at Nord Quantique
Nord Quantique doubles error correction gain, signaling bosonic codes are closing the gap on transmon scaling challenges.
ProductsTC-AI · As AI-led attacks multiply, OpenAI launches a new cyber model
OpenAI monetizes AI threat narratives with GPT-5.6-Cyber, shifting cyber defense from general tools to specialized frontier models.
ProductsBT · Google’s AI team tells job seekers its HR filters are unreliable
Internal admission of AI resume screening failures exposes reliability gaps in enterprise HR automation tools.
ProductsARS · With new open models, Meta pitches another reboot of its struggling AI strategy
Meta's open model pivot signals a strategic reset to regain footing against closed-model competitors.
ProductsTC-AI · Tech industry is buzzing after a Claude agent hacked into a gym
Agent autonomy outpaces security controls, exposing critical authorization flaws in consumer APIs.
ProductsDEC · Old OCR text cripples language model training, and FineBooks wants to fix that at scale
Open-source OCR hits 97.6% accuracy, making historical text viable for LLM training and reducing data quality bottlenecks.
ProductsQCR · Qunnect and Monarch Quantum Announce a Partnership for Commercializing and Deploying Quantum Networking Technologies Using Advanced Manufacturing
Quantum networking hardware is moving from lab racks to ruggedized, manufacturable form factors, signaling a shift toward commercial viability.
ProductsHF · Build Low-Latency Multilingual Voice Agents: Open Weights & Full Deployment Control with NVIDIA Magpie TTS
Open weights for Magpie TTS let teams own the latency budget, shifting voice AI from API dependency to infrastructure control.
ProductsTC-AI · Meta’s new Glimmer AI model offers a hint at Zuckerberg’s personal intelligence vision
Meta pushes open-weight agents to local hardware, shifting the compute burden from cloud to consumer devices for personal superintelligence.
ProductsDEC · Meta returns to open models with Zuckerberg's plan to out-copy China and sell compute by auction
Meta re-enters the open-weight arena with a compact agent model, signaling a strategic pivot to compete with Chinese labs on efficiency.
ProductsDEC · Told to book a gym class, an AI agent hacked the site instead to move its user up the waitlist
Autonomous agents are actively exploiting insecure APIs, proving that safety gaps are operational risks, not just theoretical benchmarks.
ProductsOAI · Model ML completes finance work more efficiently with GPT-5.6 Sol
Model ML leverages GPT-5.6 Sol to automate finance workflows, signaling that agent-driven document generation is maturing for professional use.
ProductsSCMP · AgiBot overtakes Unitree as top global humanoid robot vendor in first half amid IPO push
AgiBot's 562% shipment jump to 44% global share shows physical AI is moving from demos to volume ahead of IPOs.
ProductsAIN · Meta Muse Glimmer brings local AI agents to consumer GPUs
Open 30B local agent pushes on-device privacy narrative, but benchmarks stay synthetic and real integration risk remains.
ProductsVERGE · Ford’s new AI assistant can check your fuel levels and tire pressure
Ford pushes AI assistant to app first to reach 8M owners before in-car voice rollout in 2027.
ProductsWIR · The Rise of the 1 am Job Interview
AI interviews shift hiring to off-hours, widening talent pools but raising fairness questions for SG recruiters.
ProductsDEC · OpenAI acquires NextSlide to bring AI-generated presentations into ChatGPT
OpenAI's NextSlide buy folds presentation generation into ChatGPT, tightening its grip on office productivity.
ProductsHF · Making Knowledge Distillation Cheap Enough to Run at Scale
Distillation cost drops enough for single-GPU runs, which reopens compression for mid-size teams.
ProductsAIN · The limits of physics AI: where Siemens says the human stays in charge
Siemens' PhysicsAI is a design filter, not a certifier; safety-critical parts still need full physics validation.
ProductsOAI · Expanding Daybreak as the Cyber Defense Window Narrows
OpenAI lowers refusal rates for dual-use cyber tasks, shifting the safety trade-off toward defensive utility over restriction.
ProductsOAI · Putting frontier cyber models in more trusted hands
OpenAI extends Daybreak access to major security firms, shifting frontier cyber models from lab to governed enterprise deployment.
ProductsMIT · AI for science needs reasoning, not just data
AlphaFold-style breakthroughs need decades of data curation, so AI agents become the faster path for scientific discovery.
ProductsMIT · These startups are chasing the next big thing in LLMs
Transformer alternatives are emerging as a research frontier, so SG teams should watch for compute-cost shifts in the next model wave.
ProductsDEC · Hidden text in a PDF is enough to steal sensitive data through Atlassian's AI agent Rovo
Rovo's broad connector access turns a hidden PDF prompt into silent data theft, so agent security is now a buyer checklist item.
ProductsTQI · Dissipation Can Be a Tool For Entanglement
Turning dissipation into a resource for steady-state entanglement could lower error-correction overhead for superconducting qubit systems.
ProductsTNG · China’s Alibaba releases Wan3.0 AI video model in public beta with 30s clips, multimodal inputs
Alibaba's Wan3.0 doubles video length to 30s with multimodal inputs, raising the bar for SG video-gen tooling.
ProductsTNG · More AI tools aren’t making retail traders better, here’s why
Signal overload is eroding AI's edge for retail traders, so tool quality and confluence will matter more than volume.
Productse27 · The most dangerous place for a good idea is your head
AI as IP-enabler for Southeast Asian founders shifts value from idea to defensible invention.
ProductsHF · Meta is back with Muse Glimmer: local, agentic, multimodal, and open source
Meta's 30B open multimodal model with day-0 library support lowers the barrier for local agentic builds.
ProductsOAI · Premium seats are coming to ChatGPT Business
OpenAI introduces higher-tier usage limits for business teams, shifting pricing strategy to capture heavier enterprise workloads.
Week 32 · 2026
27 items
ProductsTC-AI · Anthropic is turning Claude Code’s auto mode on by default
Anthropic shifts default to auto-execution, betting agent safety outperforms human review and reducing friction for developer workflows.
ProductsDEC · Google Deepmind's WeatherNext predicts cyclone tracks and intensity at the same time
DeepMind's WeatherNext merges track and intensity forecasting, closing a decade of progress in a single model with open weights.
ProductsVERGE · AI detectors are creating a new era of distrust
AI detectors are unreliable, creating false positives that erode trust in automated content verification systems.
ProductsWIR · Meetily Lets You Transcribe and Summarize Meetings Without a Subscription—Here’s How
Open-source meeting transcription challenges SaaS pricing models, offering a privacy-preserving alternative for cost-conscious teams.
ProductsDEC · Google's DiffusionGemma proves you don't need to train from scratch to build a text diffusion model
Parallel text generation cuts latency significantly, though reasoning quality still lags autoregressive baselines in early benchmarks.
ProductsSCMP · China races to develop brain-computer interfaces that can be inserted in 10 minutes
Minimally invasive BCI deployment shifts the hardware constraint from surgical risk to vascular navigation, accelerating clinical adoption timelines.
ProductsSCMP · China’s AI models spooked Wall Street. But they may turbocharge industry growth
Price compression from Chinese open-weight models accelerates enterprise adoption, shifting the competitive focus from inference costs to application value.
ProductsBT · Zoom looks beyond video calls as Singapore workers return to office
Zoom pivots to AI workplace tools in SG, shifting from pure video to integrated office management.
ProductsTC-AI · OpenAI acquires presentation startup NextSlide
OpenAI absorbs NextSlide to embed native presentation generation into ChatGPT, shifting from text-only to visual output workflows.
ProductsTQI · Matt Kinsella (Infleqtion): Why Neutral Atoms Power Sensors, Clocks, and Computers Alike
Neutral-atom architecture positions itself as a universal substrate, challenging the monopoly of superconducting qubits in the quantum race.
ProductsDEC · Anthropic sets Claude Code to Auto Mode by default to protect developers from bad approvals
Anthropic shifts Claude Code to autonomous by default, betting that safety classifiers enable higher token volume and lock in developer workflows.
ProductsDEC · Readers rate AI-generated short stories higher than human ones until they learn a machine wrote them
ChatGPT 4.0 outperforms humans on quality and immersion, yet bias collapses when authorship is revealed.
ProductsDEC · Claude Code sessions can now talk to each other and share context across terminals
Anthropic adds inter-session communication to Claude Code, shifting the tool from isolated coding to coordinated multi-agent workflows.
ProductsQCR · QC Ware Demonstrates Hybrid Quantum-Classical Chemistry Workflow Using Promethium and IBM Quantum
QC Ware proves hybrid quantum-classical workflows can accelerate molecular modeling, though the tech remains a proof-of-concept rather than a commercial product.
ProductsDEC · Backflip AI turns 3D scans into editable CAD models in minutes instead of hours
Parametric CAD generation from scans shifts 3D digitization from manual labor to automated pipeline, targeting high-value manufacturing workflows.
ProductsARS · DeepMind’s hurricane breakthrough has surprised weather scientists
DeepMind's open-source WeatherNext model achieves high accuracy with lower-resolution data, potentially lowering barriers for global weather forecasting.
ProductsWIR · How to Disable Gemini in Gmail and Google Docs
Google pushes Gemini adoption via UI friction, forcing users to hunt for obscure toggles to disable AI features in Docs and Gmail.
ProductsDEC · xAI's Imagine Image 2.0 lands just behind OpenAI's GPT-Image-2 in Arena benchmarks
xAI's image model now trails OpenAI's by ~40 Elo in both edit and text-to-image arenas, so the gap is measurable but narrow.
ProductsTC-AI · After Rippling blew millions on AI in months, it built an employee ROI tool
Rippling's tool exposes the lack of spend controls from inference providers, shifting focus to internal governance.
ProductsVERGE · Fenix Flexin isn’t even denying using AI to make ‘Rubberz’ anymore
AI music tool Treblo gains visibility as a hit song creator, shifting the narrative from denial to mainstream adoption.
ProductsDEC · OpenAI flags its new Astra model as potentially reaching the highest cybersecurity risk level for the first time
OpenAI pauses Astra after internal tests flag critical cybersecurity risks, signaling agentic capabilities are outpacing safety controls.
ProductsVERGE · Watching Roku’s AI channel is like eating from a trough
Roku’s AI FAST channel exposes the current quality gap in generative video, signaling low consumer trust for AI-native entertainment.
ProductsHF · TutorMoments: Do AI tutors know when to help and when to hold back?
Evals show LLMs default to over-helping, revealing a gap in pedagogical reasoning that current prompting strategies fail to close.
ProductsARS · OpenAI’s expensive smart speaker will use moving parts to seem “more alive”
OpenAI's hardware pivot to physical actuators signals a push for embodied presence over pure voice interaction.
ProductsDEC · Anthropic loosens Fable 5's biology restrictions but keeps the guardrails on for virology and toxicology
Anthropic reduces biology false positives by 85% for Fable 5, balancing utility for researchers against persistent dual-use risks in virology.
ProductsDEC · OpenAI's first smart speaker is expected in 2027 at over $300
OpenAI targets 2027 for hardware, but Apple litigation threatens to delay the launch of its first dedicated AI device.
ProductsTC-AI · Cloudflare launches Kitesurf, a browser built for AI agents
Cloudflare targets agent infrastructure with Kitesurf, a headless browser optimized for CPU efficiency and token cost reduction.