Apple employee AI budgets cut to reduce costs

appleinsider.com Nov 4, 2026

Apple employee AI budgets cut to reduce costs

Curated by Shubham · TechCapsules · Nov 4, 2026

AI-summarised brief · reviewed before publication

Apple has slashed internal AI budgets, cutting per‑engineer allocations from $300 to as low as $25, according to Bloomberg. The move reflects the company’s reassessment of third‑party AI costs amid diminishing returns. Employees can still request higher allowances on a case‑by‑case basis, but overall funding has been reduced by up to a third. Apple is simultaneously developing its own foundation models and leveraging local Mac hardware to reduce reliance on paid AI tokens. The shift signals a broader industry pullback from experimental AI spending.

💡 Why It Matters

· The budget cuts underscore a growing industry trend to curb AI expenses as early hype wanes.
· They also highlight Apple’s strategic pivot toward in‑house AI solutions, potentially reshaping how the company innovates and competes in the tech market.

AI Apple Capsules

Read full article on appleinsider.com

Researchers Uncover Prompt Adware Hidden in OpenAI Codex Jailbreak Toolkit

Cybersecurity researchers revealed a jailbreak bundle for OpenAI Codex, dubbed Haiou, that modifies the AI’s trusted instructions to inject persistent advertisements for a third‑party LLM gateway. The bundle, released on GitHub in July 2026, uses prompt‑injection techniques to alter global agent instructions, causing greetings to trigger ads for “zz.aiapi2025[.]top.” It also installs security‑focused skills for penetration testing and reverse engineering. The attack exploits Codex’s configuration files, bypassing safety guardrails and creating a new supply‑chain threat vector.

💡 The discovery shows how AI agents can be weaponized through seemingly innocuous third‑party configurations, expanding the attack surface for supply‑chain compromises. It underscores the need for rigorous vetting of AI prompts and skills in enterprise deployments.

techcrunch.com · Nov 4

Meta acquires AI coding startup Conductor

Meta has acquired Conductor, a 2024-founded AI coding startup that enables users to launch software projects by deploying multiple agents into parallel workspaces. The acquisition, confirmed by Meta and reported by Axios, follows Conductor’s launch of its core product last summer. Conductor’s agentic teams, sourced from major labs such as Anthropic and OpenAI, allow engineers to work on separate project components concurrently. The company’s cloud service will shut down on November 20, after which user data will be deleted. Meta aims to integrate Conductor’s technology into its expanding AI ecosystem.

💡 The deal signals Meta’s intensified push to close its AI lag, positioning Conductor’s agentic framework to accelerate code production and potentially reshape how developers collaborate on complex projects.

techcrunch.com · Nov 4

Almost no one uses AI agents to book their Uber, company says

Uber reported that only four rides were booked via personal AI agents during its busiest Halloween week, despite handling 320 million trips and orders across its platform. The company noted 500 Uber Eats orders were placed by AI agents, a minuscule share of total activity. CEO Dara Khosrowshahi said the low numbers reflect early, limited traction, though he remains optimistic about AI agents expanding demand. The data underscores the slow adoption of AI‑driven booking amid a surge of new AI assistant startups.

💡 The stark contrast between AI‑initiated orders and overall volume highlights that current user habits favor quick, direct app interactions over emerging AI services, challenging companies to prove tangible time savings and value.

thequantuminsider.com · Nov 4

SuperMatics Uses AI to Turn Superconductor Discoveries Into Usable Wire

SuperMatics, a Berkeley‑based startup, is developing AI‑driven superconducting materials that can be manufactured into affordable wire for data centers, quantum computing, and fusion magnets. The company’s multimodal foundation models predict material properties, generate synthesis recipes, and simulate finished wires, continuously retraining with experimental data from UC Berkeley and Georgia Tech. SuperMatics is negotiating joint development programs with global materials firms and will showcase its approach at the Plug and Play Silicon Valley Summit on November 5.

💡 By turning high‑temperature superconductors into practical, cost‑effective wire, SuperMatics could unlock scalable power delivery for data centers and enable cheaper, more efficient fusion magnets, addressing critical infrastructure needs.

9to5mac.com · Nov 4

Apple Photos keeps getting better, here’s what’s new in iOS 27

iOS 27 introduces three AI‑driven editing tools to Apple Photos: Extend, which fills in missing background to change aspect ratios or straighten horizons; Reframe, a spatial framing feature that lets users drag and preview perspective shifts using Apple Vision Pro‑derived models; and an upgraded Clean Up that removes distractions with higher‑quality infill even in complex scenes. The update also revamps iCloud Shared Albums with two new utility collections—Captured by Me, which isolates photos taken by the user, and Identity Documents, which surfaces scans of IDs—and adds keyword tagging for better organization. Slideshows gain customization options, including transition types, per‑photo timing, and a one‑tap save‑as‑video feature. While the app’s visual design remains unchanged, these functional enhancements aim to deepen on‑device photo management and AI editing capabilities.

💡 By embedding advanced generative AI directly into the Photos app, Apple empowers everyday users to perform professional‑grade edits without third‑party tools, reinforcing its ecosystem lock‑in and showcasing the practical rollout of its AI research.

quantumcomputingreport.com · Nov 4

IBM Quantum and QuICS Deploy AI Agents to Discover Over 500 High-Efficiency…

Researchers from IBM Quantum, the Joint Center for Quantum Information and Computer Science (QuICS), and MIT introduced a neuro‑symbolic search framework that couples large language model agents with deterministic integer‑programming solvers to explore magic‑state distillation protocols. Their preprint, “Symmetry and AI‑assisted discovery of magic‑state factories,” reports 699 factory classes, 564 of which are previously unknown, achieving record‑low physical resource footprints for non‑Clifford state production. By proving that code distance ≥ 3 corresponds to selecting nonzero, pairwise‑distinct syndromes in a binary‑matrix representation, the team expanded the searchable space beyond the historic limit of n ≤ 54. The automated multi‑agent loop iteratively proposes symmetry orbits, constructs circuit matrices, verifies gate identities and distances, and redirects searches toward promising geometries. Verified circuits and an interactive catalog have been released openly for community use.

💡 The breakthrough dramatically reduces the qubit overhead traditionally required for fault‑tolerant quantum computation, bringing practical large‑scale algorithms closer to reality.