Tech news · 18 November 2026

Tech news, 18 November 2026: the Optimizely supply chain attack, Nvidia's first results without Huang, New York's AI enforcement unit and Microsoft's MAI Code 2

Attackers used a flaw in Optimizely's experimentation platform to inject malicious JavaScript into the websites of 178 organisations, a reminder that every third-party script runs with your site's full trust. Nvidia reported strong results and raised guidance in its first quarter under an interim chief executive, New York set up a unit to enforce its AI safety law, and Microsoft released cheaper coding models.

Tech newsSumanta Mukhopadhyay6 min read6 sources
#news#security#ai#governance
TECH NEWS · TECH NEWS · 18 NOVEMBER 202618 Nov 2026 · 08:50 GST

Today's brief covers stories from 16 to 17 November.

  1. Sign upAn attacker abuses a flaw in a vendor's account sign-up flow
  2. WriteThey gain write access to other customers' projects
  3. InjectMalicious JavaScript is added to a custom code field
  4. ServeThe vendor's CDN delivers it to every page of customers' sites
  5. StealVisitors' payment details and crypto transfers are targeted
  6. ContainThe vendor fixes the flaw, removes code and revokes tokens
How a flaw in one vendor became malware on 178 organisations' websites, as reported.

1. Attackers inject malware into 178 websites through Optimizely

Optimizely said an unauthorised party exploited a flaw in its Experimentation account sign-up flow to inject malicious JavaScript into some customer projects; Report URI identified infected snippets across 178 organisations, including major media and technology brands1. The Register reports the code targeted visitors' payment details and cryptocurrency transfers, and Optimizely says it fixed the flaw, removed the code, revoked tokens and notified customers2.

CTO take: any third-party script on your site runs with your site's full trust. List every script you load, keep testing, analytics and marketing tags off payment and login pages, restrict what scripts can do with a Content Security Policy, monitor scripts for unexpected changes and make sure you can switch a vendor's tag off in minutes. Payment card rules already require monitoring of scripts on payment pages; this shows why. See the payment page script security case study.

2. Nvidia beats and raises in its first results without Huang

Nvidia reported third-quarter fiscal 2027 revenue of $115.8 billion, with data centre revenue of $107.6 billion, and guided next-quarter revenue to about $130 billion3. CNBC notes it was the company's first report since Jensen Huang resigned as chief executive, with Colette Kress serving as interim CEO4.

CTO take: demand for AI hardware is still rising, which means capacity stays tight and expensive for most buyers into next year. Plan for that: commit to capacity only where usage is proven, make models cheaper to run, and keep your stack portable across providers. See AI capacity contracts.

3. New York creates a unit to enforce its AI safety law

New York Attorney General Letitia James announced an AI Safety, Standards, Enforcement and Transparency unit, which her office describes as the first in the country dedicated to enforcing AI laws, and which will enforce the state's RAISE Act for frontier AI developers5.

CTO take: enforcement capacity is following the laws. The direct duties fall on large model developers, but expect providers to ask customers for help with incident reporting and to update terms. Keep a record of the AI models you use and how; see the board AI briefing checklist and the US page.

4. Microsoft releases MAI Code 2 for coding

Microsoft announced MAI Code 2 and MAI Code 2 Flash at Ignite, available in GitHub Copilot, alongside other new MAI models; Microsoft's own benchmarks claim performance close to leading frontier models at a much lower price6.

CTO take: coding model choice is becoming a cost lever, not just a quality one. Test new models on your own repositories and tasks before switching, and keep your coding agent policy independent of any one model; see AI coding policy for engineering teams.

What I'd do this week

  • Inventory third-party scripts on your sites and remove any from payment and login pages.
  • Add or tighten a Content Security Policy and script change monitoring.
  • Revisit AI capacity plans in light of continued tight supply.
  • Trial new coding models on real tasks before changing defaults.

I am based in India, the United States and the UAE, working globally. See where I work, or analyse your challenge and bring the result to a call.

References
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
FAQ

Common questions

What happened in the Optimizely attack?

Optimizely said an unauthorised party exploited a flaw in its Experimentation account sign-up flow to inject malicious JavaScript into some customer projects. Report URI identified infected snippets across 178 organisations. Optimizely says it fixed the flaw, removed the code, revoked tokens and notified customers.

How can websites protect themselves from compromised third-party scripts?

Inventory every third-party script, limit what each can do with a Content Security Policy, keep scripts off payment and login pages where possible, monitor scripts for unexpected changes and be able to switch a vendor's script off quickly.

What is New York's ASSET unit?

A new unit in the New York Attorney General's office dedicated to AI safety, standards, enforcement and transparency, which will enforce the state's RAISE Act for frontier AI developers when it takes effect.

Keep reading

Related

ServiceAI & LLM architectureI design and build AI systems that leadership can trust: grounded in your data, measured by evaluations, governed by your rules and priced so they make sense at scale.→ServiceFractional CTO and AI architect for US companiesSenior technology leadership for US founders, enterprises and investors: AI that reaches production, data platforms that pass security review, and governance that keeps pace with federal and state rules.→InsightAI inference providers are now supply chain: what the NEAR AI and Fireworks breaches teach buyersWithin days of each other, NEAR AI and Fireworks AI disclosed breaches. One came through a flaw in an open-source inference engine running inside a confidential computing enclave; the other through a cloud provider hosting research machines, reaching customers' secrets, datasets and partial model weights. Most companies now send prompts, documents and fine-tuning data to providers like these. Here is what to learn, and what to ask.→InsightTech news, 11 November 2026: Australia's frontier AI rules, an exploited Outlook zero-day, Fortinet's advisories, New York's AI safety law and Anthropic's office toolsAustralia announced it will legislate mandatory safety rules for frontier AI developers, policed by a new regulator. Patch Tuesday brought an exploited Outlook flaw, Fortinet published details of critical fixes it shipped in October, New York set registration rules for frontier developers and Anthropic moved into office software.→
Open to advisory, fractional CTO & select builds

Let's talk about what you're building.

The fastest route is a free 30-minute call. Prefer to write? Send a message and I'll reply personally within two business days.

Start a conversation

How would you like to reach me?

Book a 30-min callFastest. Pick any open slot on my calendar.→ Write in GmailOpens a pre-filled draft in your browser.→ Use my email appOutlook, Apple Mail or your default client.→

I reply personally within two business days. Time zones from the Americas and Europe to the Middle East and Asia-Pacific.