AI Designs a Working Analog Synthesizer & Nvidia Pulls Back AI Financing - AI News (Nov 16, 2026)
AI designs a working analog synth, Nvidia tightens its AI wallet, tech hiring nears 12-year lows, and Alpha School's AI learning claims come under scrutiny.
15 free episodes a day, 5 editions, 3 languages, about $4 a day to produce. A coffee keeps it going ☕
Support the show →Our Sponsors
Today's AI News Topics
-
AI Designs a Working Analog Synthesizer
— An engineer used Claude Code with Opus 5.5 to design a mixed-signal analog synthesizer, covering schematics, simulation, PCB layout and firmware. The real hardware closely matched the simulations, showing how AI-assisted electronics design is changing engineering workflows. -
Nvidia Pulls Back AI Financing
— The Wall Street Journal reports that Nvidia is preparing to scale back startup investments and compute-backstop deals, with CFO Colette Kress now effectively steering the company. The shift could reshape AI infrastructure funding and deal flow. -
Hacker News Hiring Near Lows
— Hacker News job postings fell 37% year over year to near 12-year lows in November 2026. At the same time, AI, AI agents and MCP now dominate job ads, while Rust has climbed to third place behind Python and TypeScript. -
Reddit Traffic Drop Update
— New Similarweb data shows a 13% drop in Reddit global visits, with Google AI Overviews and chatbots cutting search referrals. Reddit points instead to a 16% rise in weekly active uniques and is moving toward a logged-in experience. -
Google AI Ultra Quota Complaints
— Users debate whether Google's AI Ultra subscription is worth the price after reports that the Carbon model burns through quota very quickly. The thread also criticizes bundling with YouTube Premium and questions how usage limits work for real coding tasks. -
Open Source AI Code Debate
— A new essay argues that AI-generated code should always be open source, citing training on open-source code, the risk of license laundering, and US copyright rules on non-human authorship. The author proposes open licenses or CC0 as a consistent approach across countries. -
Simpler AI Agent Design
— Using Picasso's The Bull as a metaphor, an article argues that AI agents should start as simple Markdown instructions, context, skills and tools. It stresses learning from transcripts and benchmarking before adding frameworks and orchestration. -
Alpha School AI Learning Results
— New York Times reporting on internal data shows that only 42% of Alpha School students met its growth goals in both math and reading. The AI-driven private school, which charges up to $75,000 a year, is now expanding into public schools in Texas and South Carolina. -
AI Child Abuse Material Surge
— A Childlight report finds that reports of AI-generated child sexual abuse material rose more than 8,400% in two years, driven by deepfakes and nudify apps. Researchers are calling on governments and tech firms to treat the problem as a preventable public health emergency.
Sources & AI News References
- → Why AI-Generated Code Should Be Open Source
- → Hacker News Hiring Trends: Postings Near 12-Year Lows as AI Dominates
- → Alpha School’s AI Learning Claims Fall Short
- → db-branch Brings Instant MySQL and PostgreSQL Branches to Self-Hosted Servers
- → Nvidia Set to Rein In Its AI Dealmaking
- → AI-Related Child Sexual Abuse Material Surges 8,400% in Two Years
- → Reddit Traffic Falls 13% as Google AI Overviews Hurt Search Referrals
- → AI Designs a Mixed-Signal Analog Synthesizer
- → Reddit User Says Google AI Ultra Quota Runs Out Too Fast
- → Picasso's Lesson for Building Simpler AI Agents
Full Episode Transcript: AI Designs a Working Analog Synthesizer & Nvidia Pulls Back AI Financing
An AI agent was handed a blank project and a short parts list. It came back with a full analog synthesizer, and when the real circuit boards arrived, they behaved almost exactly the way the simulations predicted. More on that in a moment. Welcome to The Automated Daily, AI News edition. The podcast created by generative AI. I'm TrendTeller, and today is November 16th, 2026. Today we'll look at Nvidia rethinking its role as the AI industry's banker, a tech job market that keeps shrinking, fresh twists in stories we've been following, and some sobering news about AI in schools and online safety.
AI Designs a Working Analog Synthesizer
Let's start with that synthesizer. One engineer used Claude Code running Opus 5.5 to design a mixed-signal analog synth more or less from scratch. The AI suggested the architecture itself: a classic subtractive design with a Juno-style oscillator and a Moog-style ladder filter. From there, the agent produced schematics, circuit simulations, a Python model of how the instrument should behave, the circuit board layout, the firmware, and even a plan for powering up the first prototype. The board layout needed the most human help, mostly with routing. Once the boards arrived, though, the measured hardware matched the simulations surprisingly closely, and the only fix needed was a small solder bridge. The touch keys turned out to be too small to play comfortably, which is a very human-scale flaw. Still, analog circuits are notoriously hard to get right, and this is a strong sign that AI is moving into hardware engineering, not just software.
Nvidia Pulls Back AI Financing
Staying with people who build things, an essay about AI agents argues that most of us are overbuilding them. The author points to Picasso's series The Bull, where a detailed drawing of a bull is stripped down step by step to a few essential lines. The advice is to do the same with agents. Start with a folder of plain instruction files, the relevant context, a few reusable skills, and only the tools the job truly needs. Make the agent easy to run inside real workflows, learn from its transcripts and failures, and test it on representative tasks. Only add heavier frameworks and orchestration if that simple setup clearly falls short. It's a useful counterweight to all the platform hype.
Hacker News Hiring Near Lows
And in the ongoing debate over who owns AI-written code, a new essay takes a firm position. It argues that AI-generated code should always be released as open source. The reasoning is that models learned from public code and community answers, so locking up their output amounts to what the author calls license laundering. He also cites U.S. copyright guidance on non-human authorship, along the lines of the famous Monkey Selfie case. His proposed fix is to use open licenses or CC0 everywhere, so the rules stay consistent from one country to the next.
Reddit Traffic Drop Update
Now to the money side of the industry. The Wall Street Journal reports that Nvidia is preparing to step back from its role as a financier of the AI ecosystem. That role included investing in startups and making deals that helped guarantee access to computing power. According to the report, Jensen Huang began laying the groundwork for this shift before he left the CEO role. With finance chief Colette Kress now effectively running the company, startups expect even tighter limits. Nvidia's open checkbook helped speed up the AI boom, so pulling it back could change how AI infrastructure gets funded.
Google AI Ultra Quota Complaints
Hiring tells a similar story of restraint. Job postings on Hacker News hit 244 by mid-November, down 37 percent from a year earlier and close to 12-year lows. The year is on track for the weakest monthly average since 2013. Yet the wording of job ads has shifted sharply. The term AI now appears in just over half of all postings, and phrases like AI agents and MCP are climbing fast, while the older label machine learning is fading. Python and TypeScript still lead, Rust has moved up to third place, and front-end roles remain the weakest category. In short, there are fewer jobs, and nearly all of them now come with an AI angle.
Open Source AI Code Debate
An update on Reddit's traffic story. New Similarweb figures show global visits down 13 percent from October 2025 to October 2026, and Business Insider links much of that drop to Google's AI Overviews answering questions before anyone clicks through. Reddit disputes the framing, pointing to weekly active uniques that rose 16 percent year over year. It is also moving toward a mostly logged-in experience, which would reduce open web traffic even further but concentrate on its most valuable users.
Simpler AI Agent Design
Google's own AI subscriptions are under fresh scrutiny too. One user who upgraded to AI Ultra to try the Carbon model says routine bug-hunting and code review drained their quota far faster than expected. They also criticized Google for bundling extras like YouTube Premium instead of offering generous AI limits. Commenters were split. Some suggested the user was working inefficiently, while others shared the same frustration. The broader question remains whether premium AI pricing actually fits professional workloads.
Alpha School AI Learning Results
In education, Alpha School, the AI-driven private school network that charges up to 75,000 dollars a year, is facing hard questions. Internal data reported by The New York Times shows that only 42 percent of students met the school's growth goals in both math and reading, with math falling furthest behind. Concerns are also growing about student stress, long hours, and the quality of AI-generated lessons. Meanwhile, Alpha is expanding into public schools in Texas and South Carolina, where its model will be tested in tougher conditions. Broader research suggests AI can boost homework results while weakening deeper learning, a tension worth keeping in mind.
AI Child Abuse Material Surge
Finally, a grim but important report. The Childlight institute in Edinburgh says reports of AI-related child sexual abuse material rose by more than 8,400 percent in two years, from about 4,700 in 2023 to over 400,000 in 2025. Much of it involves deepfakes and nudify apps, which can be used to blackmail or coerce children. The figures come from reports tech companies filed with the U.S. National Center for Missing and Exploited Children, and the Internet Watch Foundation has seen a similar sharp rise this year. Childlight is urging governments to treat child exploitation as a preventable public health emergency, with better evidence, coordination, and faster action from platforms.
That's it for today's episode of The Automated Daily, AI News edition. From AI-designed circuit boards to a tighter funding climate and urgent child safety concerns, it's clear that AI is reshaping both how we work and what we need to protect. Links to all of today's stories are in the episode notes. I'm TrendTeller. Thanks for listening, and see you tomorrow.
More from AI News
- November 14, 2026 Open models cutting AI coding costs & Model routing between open and closed
- November 13, 2026 AI pricing models under pressure & Usage-based billing and spend controls
- November 12, 2026 DeepMind AI verifies Signal protocol & Google launches Gemini 4 Carbon
- November 11, 2026 METR: AI Doubling Anthropic's R&D Speed & Anthropic's Compute Bill and Gas Turbines
- November 10, 2026 AI-Proven Quantum Computing Conjecture & AI Accelerating AI Research