AI headlines are getting harder to categorize — part finance, part hardware, part ethics seminar. This week’s mix captures that split personality: money pouring in at the top, chips doing the heavy lifting underneath, and policy trying to keep up somewhere in the middle.

LLM Expert Insights,

Packt

Marketrebalanceand strategicbets

Nomura sees modest IT rebound but warns AI hype is still holding things back

Nomura expects a slightly better year ahead for India’s IT sector, forecasting around 4.5% dollar-revenue growth for large-cap firms and a modest margin boost helped by favorable currency trends. At the same time, the firm notes that many global investors still view traditional IT-services companies as “AI losers”,aperceptionclouding confidence even as companies adapt.(Moneycontrol)

Bill Gates’s daughter lands $30 million for AI-shopping startup

Phoebe Gates, 23, has raised $30 million for her AI-driven shopping assistant startup Phia,a leap that values the company at around $180 million just months after its seed round. Phia aims to use AI to simplify online shopping: comparing prices across thousands of retailers, surfacing resale options, andconsolidatingdiscounts and deals for users. The round attracted heavyweight backers (celebs and VCs alike), spotlighting how investors are placing big bets on “shopping-agent” style AI tools to redefine retail.(Bloomberg)

Infrastructureandchipsunderpinning AIgrowth

Micron pulls Crucial from stores as AI memory demand peaks

Micron is phasing out its “Crucial” line of consumer-grade RAM and SSD products by February 2026 —it’sredirecting all production capacity toward high-bandwidth memory (HBM) and enterprise-grade chips used in AI data centers. The decision reflects how soaring demand for AI infrastructure is reshaping the memory-chip market: what used to be widely available to PC builders and everyday users is now being reservedalmost exclusivelyfor AI-backed workloads and large-scale data-center customers.(micron.com)

Nvidia doubles down on chip-design upgrades with$2 billionSynopsys investment

Nvidia has committed$2 billionto Synopsys, deepening a multi-year partnership aimed at embedding AI and GPU-accelerated computing across the chip-design and product-engineering stack. Under the deal, Synopsys will integrate Nvidia’s AI toolsets,including CUDA libraries, physics-powered AI workflows, and digital-twin simulation platforms,to speed up design cycles from individual chips to entire systems in industries ranging from semiconductors to aerospace and automotive. The move signals that AI’s next big frontierisn’tjust model training or inferencebut alsothe hidden infrastructure behind every smart product.(Nvidia)

Regulation, Risk & AI Ethics in Deployment

Google’s Nano Banana Pro sparks outrage for racial bias in AI-generated imagery

Google’s new image-generation tool, Nano Banana Pro, has come under fire after users found it repeatedly depicting “whitesaviour” scenes,such as white volunteers surrounded by Black children,in response to prompts about humanitarian work in Africa. The model also inserted real NGO logos like Save the Children and World Vision without consent, raising fresh concerns about bias, representation, and misuse of brand identity in AI outputs.(The Guardian)

EU launches antitrust probe into Meta over WhatsApp’s AI policy

Meta is under fresh regulatory scrutiny in Europe after the European Commission opened a formal antitrust investigation into WhatsApp’s new policy around AI chatbots. The concern: Meta may be blocking rival AI providers from offering chatbot services via WhatsApp, while giving its own Meta AI a privileged spot,potentially tilting competition in itsfavour.(Reuters)

India turns to AI-powered surveillance to curb exam cheating

India’s national education boards are rolling out AI-enabled CCTV cameras across testcentresto prevent leaks and malpractice,starting with pilot programs intheeastern stateofOdisha. The system uses real-time monitoring and alerting to flag irregularities during both written and practical exams, marking one of the country’s first large-scale experiments in automated exam proctoring.(Times of India)

First-of-its-kind AI-orchestrated cyberattack exposes new risks for global security

The firm Anthropic says it disruptedwhat appears to be thefirst large-scale cyber-espionage campaignlargely automatedby AI,using its coding-agent tool Claude Code to run reconnaissance, exploit vulnerabilities, and harvest data acrossroughly 30globalorganisations. According to the disclosure, the AI performed 80–90% of the tactical operations with minimal human direction, blurring the line between “smart assistant” and “automated attacker.” The incident marks a watershed in cybersecurity,showing that AI tools meant to boost productivity can be turned into weapons byactorssavvy enough to hijack them.(AI News)

Taken together, these stories make one thing clear:AIisn’tone industryanymore,it’sseveral competing realities. And depending on which oneyou’rewatching, the future looks like a jackpotor a legal headache.

Best part? They’re running their Holiday Season Giveaway and first 100 people get in for absolutely free (it usually costs $395 to attend)

Live sessions- Saturday and Sunday
10 AM EST to 7PM EST

From justresearch experiments, agentic systems have evolved into practical, high-impact solutions that enterprises now depend on. But as anyteamtasked with deploying AI solutionsquickly discovers, building these systems from scratch is complex. Different agents need roles, tools, memory, state, and reasoning loops,and they must work together reliably and transparently.

InAgentic Architectural Patterns for Building Multi-Agent Systems, authorsDr. AliArsanjaniandJuan Pablo Bustosaddressthis challengedirectly. Their chapter on Agent Frameworks moves beyond theoryand showshow modern tooling,specificallyCrewAIandLangGraph,enables developers to translate agentic principles into production-ready applications.

This week’sAI_DistilledExpert Insightexplores a real, enterprise-grade example by reconstructing the same multi-agent loan processing system in two ways,once usingCrewAIand again usingLangGraph.This revealshow different frameworks shape the same workflow.This article offers a clear, comparative look at two major approaches to agent engineering, highlighting the strengths and trade-offs of each.

Whether it's a scrappy prototype or a production-grade agent, we want to hear how you're putting generative AI to work. Drop us your story at nimishad@packtpub.com or reply to this email, and you could get featured in an upcoming issue of AI_Distilled.

📢 If your company is interested in reaching an audience of developers and, technical professionals, and decision makers, you may want toadvertise with us.

If you have any comments or feedback, just reply back to this email.

Thanks for reading and have a great day!

That’s a wrap for this week’s edition of AI_Distilled 🧠⚙️

We would love to know what you thought—your feedback helps us keep leveling up.

Thanks for reading,

The AI_Distilled Team
(Curated by humans. Powered by curiosity.)