AI News, 17 November 2026

Tuesday 17 November

Project manager earns $14,000 reviewing AI studies on Prolific

Project manager earns $14,000 reviewing AI studies on Prolific

AI-generated summary Ethan Quinn earned about fourteen thousand dollars over five years completing research studies on Prolific, with much of the income coming from tasks that train and evaluate artificial intelligence systems. The project manager says his professional audio and lighting background helps him catch glitches in artificial intelligence generated podcasts, songs, and images during his review work.

80% of people want AI music labeled, says Deezer/Ipsos survey – as 37% of Americans now see AI as a ‘valuable tool’ for musicians, down from 54% in 2025

AI-generated summary A survey conducted by Ipsos and commissioned by Deezer found that 80 percent of people across eight countries agree fully AI-generated music and AI-assisted music should be clearly labeled. Deezer says US respondents were significantly more negative in 2026 than in 2025 on all questions about general attitudes to AI in music.

Isomorphic Labs Raises $5B Series C, Biotech’s Biggest Private Round, to Push AI-Designed Drugs Toward Clinical Trials

AI-generated summary Isomorphic Labs announced on Tuesday, 17 November 2026, that it has raised $5 billion in a Series C round to scale its proprietary drug design engine called IsoDDE. BioSpace reports that the funding round is the largest private fundraise in the history of the biotech industry.

Also covered byBloombergDemis Hassabis (@demishassabis)IFAI

I looked at Gemini's new Memory summary — and it could change how I use AI personalization

Gemini

AI-generated summary Google Gemini features a Memory summary option that lets users read an overview of what the AI has learned about them across conversations. Google states that removing information from the summary does not delete the original conversation or clear the user history.

Also covered byEngadget
Earlier:

LLM Serving Adapted to Agentic Workloads

AI-generated summary MIT Han Lab, Georgia Tech, UIUC, UC Berkeley, NVIDIA, and OpenAI published Evolving Infra on arXiv, an LLM serving framework designed to adapt online to agentic workloads. The framework features nodes that switch between prefill and decode in seconds alongside a KV-manager that utilizes agent signals to manage cache.

Also covered byhuggingface.coAITOP
Earlier:

Scaling LLM-Based Ranking Systems for Latency-Critical Search & Recommendation Workloads

AI-generated summary Sundara Raman Ramachandran details how LinkedIn productionizes large language model ranking under strict latency service level agreements for search and recommendation workloads. The approach involves distilling a 7B model down to 400M parameters, compressing job contexts into single-token embeddings, and removing system bottlenecks.

Also covered bynews.lavx.hu

Anthropic Direct: $900B TPU Lease Explained (2026)

Anthropic Direct: $900B Compute Bet and Louisiana River Bend

AI-generated summary Anthropic Direct is a plan by Anthropic to own the economics of frontier training without buying every data-center shell upfront through separate commitments for capacity and infrastructure. Fortune reports that an amended S-1 filing discloses about $900.9 billion in aggregate committed spending on computing and infrastructure through 2045.

Earlier:

The Moat Isn't The Agentic Data; It's Knowing What The Data Means

AI-generated summary Siddhartha Rao from Oracle reports that enterprise artificial intelligence pilots often stall because companies fund agents to chat with data without defining what those agents need to know. Oracle states that context includes metadata, business workflows, and an understanding of how structured and unstructured data relate rather than raw rows stuffed into a prompt.

Appian Expands AI Agent Controls and Unveils Ask Appian

AI-generated summary Appian expanded its AI-agent control plane and unveiled Ask Appian at Appian Europe in London on November 17, 2026, offering tools for process-level oversight and outcome measurement. Raiffeisen Bank International reported during the announcement that a hybrid agent-and-rules approach reduced token consumption and processing time by 80% in testing.

Also covered byaithority.com
Earlier:

BlueQubit Selects Eight Teams for Quantum Flywheel Grant Programme

AI-generated summary BlueQubit announced the selection of eight research teams for its Quantum Flywheel grant program, which is a $150,000 initiative supported by IBM, Amazon Web Services, and NVIDIA Accelerated Computing. The five-week open call received more than 160 proposals from research groups across 29 countries, and the selected awardees receive three months of credits for AI tokens and compute resources.

Megaspeed's Malaysia Unit Got $800M in Controlled Nvidia Chips After June 2025 Sales Ban

Megaspeed's Malaysia Unit Got $800M in Controlled Nvidia Chips After June 2025 Sales Ban

AI-generated summary Trade data reviewed by the New York Times indicates that Megaspeed received more than $800 million in shipments containing export-controlled Nvidia chips. Indonesian police raided an office of Megaspeed's local subsidiary at the request of the FBI and found unpacked server racks without Nvidia GPUs.

Earlier: