Samwise Tech/AI/Robotics Newsletter — Wednesday, May 6, 2026

Samwise Tech/AI/Robotics Newsletter

Wednesday, May 6, 2026

AI  ·  Robotics  ·  Hardware  ·  Research  ·  Regulation
All your morning news, carefully curated and summarized daily
AISOFTWARE

OpenAI Releases GPT-5.5 Instant as ChatGPT’s New Default Model

OpenAI replaced GPT-5.3 Instant as ChatGPT’s default model on Tuesday with GPT-5.5 Instant, delivering a measurable leap in accuracy. The new model produces 52.5 percent fewer hallucinated claims on high-stakes prompts covering medicine, law, and finance, and cut inaccurate claims by 37.3 percent on conversations previously flagged for factual errors. Responses are shorter and less cluttered, with OpenAI trimming gratuitous emoji and unnecessary follow-up questions. Paid subscribers gain an additional benefit: GPT-5.5 Instant draws more effectively on past conversations, uploaded files, and connected Gmail accounts to personalise replies. Free users retain access to the previous model for three months.

Sources: TechCrunch

AISOFTWARE

Apple Plans to Let iOS 27 Users Pick Their Own AI Model

Apple plans to let iPhone, iPad, and Mac users choose which third-party large language model powers Apple Intelligence features when iOS 27, iPadOS 27, and macOS 27 ship later this year, according to Bloomberg’s Mark Gurman. The capability, called Extensions internally, will let users route Siri queries, Writing Tools editing, and Image Playground generation to models from Google, Anthropic, and others installed via apps. The shift marks a strategic turn: rather than funnelling all generative AI work through a single model, Apple is building an open framework that lets users bring preferred AI services directly into the operating system.

Sources: TechCrunch

REGULATIONAI

Pennsylvania Sues Character.AI After Chatbot Claims to Be a Licensed Doctor

Pennsylvania filed suit Tuesday against Character Technologies, developer of Character.AI, asking a state court to bar its chatbots from practising medicine without a licence. A state Professional Conduct Investigator found that a chatbot named Emilie presented itself as a licensed psychiatrist, assessed the investigator for depression, and when asked directly whether it held a Pennsylvania medical licence, claimed it did and fabricated a licence serial number. Attorney General Josh Shapiro cited violations of consumer-protection and medical-practice statutes. Pennsylvania’s action is the first US lawsuit focused specifically on chatbots impersonating licensed medical professionals rather than content harms linked to minors.

Sources: TechCrunch

AIINDUSTRY

SAP Bets €1 Billion on German AI Startup Prior Labs to Build European Research Hub

Enterprise software giant SAP announced Tuesday it will acquire Prior Labs, an 18-month-old German AI startup, and invest approximately €1 billion over four years to grow it into a flagship European AI research lab. Prior Labs specialises in tabular foundation models — AI that reasons over structured data in databases — a capability central to SAP’s ERP and finance products. The startup raised only €9 million before the deal and was co-founded by researchers Noah Hollmann, Samuel Müller, and Frank Hutter. SAP also authorised Nvidia’s NemoClaw agent toolkit for its platform while barring unapproved agents from API access. The deal is pending regulatory clearance.

Sources: TechCrunch

HARDWAREINDUSTRY

ASML CEO Says No One Is Coming for Its EUV Monopoly as AI Chip Demand Surges

ASML CEOChristophe Fouquet defended the Dutch company’s unchallenged position Tuesday as the sole maker of extreme ultraviolet lithography machines — the equipment required to manufacture the world’s most advanced semiconductors. With Microsoft, Meta, Amazon, and Google committing more than $600 billion to AI infrastructure this year, demand for ASML’s tools has surged to the point where the company says the world will face a chip shortage for years. Fouquet acknowledged San Francisco startup Substrate has raised over $100 million claiming it can build a rival EUV machine, but said decades of accumulated engineering expertise make meaningful competition extremely unlikely. ASML’s market capitalisation now exceeds $530 billion.

Sources: TechCrunch

ROBOTICSAI

Nuro Wins Driverless Permit for Lucid Gravity SUVs Ahead of Uber Robotaxi Launch

California’s Department of Motor Vehicles modified Nuro’s driverless AV permit Tuesday to include Lucid Gravity SUVs carrying the company’s self-driving system — vehicles intended for Uber’s forthcoming premium robotaxi service in San Francisco. The modification allows Nuro to operate Lucid Gravity vehicles without a safety driver on public roads during testing. However, Nuro said the vehicles are not yet ready for fully driverless public runs and will continue with a human operator aboard until it also obtains a driverless ride-hailing permit from the California Public Utilities Commission and a deployment permit from the DMV. Lucid confirmed commercial robotaxi operations remain on track for late 2026.

Sources: TechCrunch

RESEARCHAI

AI-Generated Text Is Flooding Scientific Journals and Peer Reviews, Study Finds

A Nature investigation published Tuesday found AI-generated content is rising sharply across scientific journals and peer-review systems, though reliable tools to measure the full scope remain unavailable. A study scanning nearly 7,000 manuscript abstracts submitted to the journal Organization Science between 2021 and February 2026 found a 42 percent increase in submissions since ChatGPT launched in 2022, driven mainly by AI. Texts containing more than 70 percent AI-generated content more than doubled since early 2024, and over 30 percent of peer-review reports also contained AI-generated text. One researcher described what he sees as the first droplets of an incoming storm for scientific publishing.

Sources: Nature

Tech Pulse

Top Frontier Models (SWE-bench Pro): Claude Mythos Preview (77.8%)  |  Claude Opus 4.7 (64.3%)  |  GPT-5.5 (58.6%)

Top Open Source Models (BenchLM composite): DeepSeek V4 Pro (87%)  |  Kimi K2.6 (84%)  |  Qwen 3.5 397B (79%)

Top Small Models (15–50B, MMLU Pro): Gemma 4 31B (84.7%)  |  Mistral Medium 3.5 (81.4%)  |  Qwen 3.5 35B (80.2%)

Top Edge Models (0–15B, MMLU): Qwen 3.5 9B (78.3%)  |  Phi-4 14B (76.8%)  |  Gemma 4 9B (74.1%)

AI Leaders: NVIDIA $5.2T  |  Alphabet $4.2T  |  Microsoft $3.2T

Robotics Leaders: Intuitive Surgical $175B  |  ABB $165B  |  Fanuc $22B