Fast & Broken

Today is a Carly^2 episode with Carly Ho, a software engineer-slash-paralegal. We discuss the website isaiprofitable.com and recent developments in AI economics, a postponed executive order from Trump and a real one from Gavin Newsom, and the Pope's AI encyclical and the value of having a variety of perspectives. -- Show notes: AI economics https://isaiprofitable.com   Anthropic projects Q2 operat...

Today's episode is focused on AI in law, and features Sarah Frankel, an Australian laywer working in immigration. We discuss the trend of lawyers submitting AI-generated briefs with hallucinated citations, where LLM tools can be useful for lawyers, how the problems we see in law are common across all domains, the issues that arise when you have LLMs making ethical decisions, how LLMs value facts,...

Today's show is cohosted by Oz Brown! Oz is nontraditional software engineer who went through bootcamp at Amazon after a history doing creative work, including voice acting, and customer support, including for Xbox. We talk about the philosophy and current state of agentic coding tools, the Anthropic-SpaceX deal and the state of circular funding, and the double standards in platform content policy...

This week I'm joined by Frank Kane, independent educator and one of the handful of people I know posting sane takes about AI on LinkedIn. We have the AI job loss conversation in the context of reporting that software engineer job listings are actually rising, go over recent developments in the realm of universal basic income (or similar things), and then talk about Anthropic's "Persona Selection M...

There will be a hiatus in weekly episodes of this podcast and there may be some newer things coming soon! Email podcast@graphenesoftware.com with suggestions or if you'd like to come on the pod!

We start the episode with a less-than-lightning round through some events of the week, including Block's 40% layoffs, Netflix walking away from the Warner Bros acquisition, and Anthropic's updates to their Responsible Scaling Policy. Then we discuss at length the drama around Anthropic refusing the Department of Defense's contract terms, and OpenAI coming in to accept the deal a few hours later. -...

After a lightning round of updates in government interaction with tech, we discuss the fascinating case study of an OpenClaw agent writing a call-out blog post for an open source maintainer who rejected its pull request, and 404 Media's detailed investigation of Alpha School, the AI enabled school surrounded by hype. For more on OpenClaw: https://carlyrector.substack.com/p/openclaw-mj-rathbun-and-...

We almost managed to not talk about AI today. This week's topics include the backlash to Discord rolling out age verification worldwide, SpaceX acquiring xAI and Elon Musk's new moon-based visions, and the story the success of Iron Lung tells about independent filmmaking. -- Show Notes Discord Age Verification https://discord.com/press-releases/discord-launches-teen-by-default-settings-globally ht...

In this accidentally all-Anthropic installment, we discuss Anthropic's firing shots at OpenAI with their new anti-ads ads, the AI tools blowing up this past week (Moltbook and the SaaSpocalypse), and insights into Anthropic's copyright strategy from documents recently unsealed from their settled lawsuit. -- Show Notes: Anthropic Super Bowl ad disses ads in ChatGPT  https://www.youtube.com/watch?v=...

Sign the petition to call on tech leaders to speak out against ICE and Border Patrol's actions in Minnesota: https://iceout.tech/ This week we discuss accusations of censorship by the newly-divested US TikTok, landmark lawsuits using a product safety argument against social media companies, and the trend of bad signs for OpenAI. -- Show Notes: TikTok https://usdsjv.tiktok.com https://www.nytimes.c...

This week, we talk about comments from Dario Amodei and Satya Nadella at the World Economic Forum, Razer CEO Min-Ling Tan's interview with Nilay Patel of the Verge, and Anthropic's official release of Claude's Constitution, previously referred to on this podcast as the soul doc. -- Show Notes: AI at Davos https://www.wsj.com/tech/ai/anthropic-ceo-says-government-should-help-ensure-ais-economic-ups...

In this episode, we review developments in the Grok deepfake epidemic and examine its implications on society at large, and then discuss the wave of excitement over Claude Code and the launch of Claude Cowork. -- Show Notes: Grok Updates Musk calls UK fascist https://www.bloomberg.com/news/articles/2026-01-10/musk-calls-uk-fascist-as-row-over-x-s-grok-ai-images-escalates   Senate bill https://www....

In this episode, we discuss our depression and anger at the current torrent of nonconsensual nudification generated by Grok on X, a $400k Polymarket win on Maduro's deposition and Congress's response, and ways data centers can be connected to the grid earlier by prioritizing workloads. -- Show Notes: Grok Deepfakes https://www.bloomberg.com/news/articles/2026-01-07/musk-s-grok-ai-generated-thousan...

In our first episode of the year, we talk about some 2026 outlook topics: potential IPOs from OpenAI and Anthropic, and AI backlash in the US midterm elections. In the latter part of the podcast, we take a look back at what years working in Amazon retail personalization can teach us about capitalism. -- Show Notes: Opening comments https://www.theinformation.com/articles/salesforce-executives-say-...

This week, we discuss the deal between Disney and OpenAI, including the interesting specifics of the financials and the challenge of implementing Disney's guardrails. Then, we discuss the complex issue of AI water use and how it's overblown in online discourse. Finally, we talk about how a lot of negative backlash to AI is actually backlash to decision's from leadership. At the end, we touch brief...

In this episode, we discuss research from DX and Anthropic on how much benefit enterprises are seeing from AI coding tools, the Polymarket bet on the Russia-Ukraine war that was probably gamed, and Australia's ban on under-16 users for a number of social media platforms going into effect. -- Industry AI usage reports: DX Q4 impact report: https://getdx.com/uploads/ai-assisted-engineering-q4-impact...

This week, we hit the viral discourse that everyone in Seattle hates AI, new research on how social media effects emotional affect and political polarization, and talk about Claude 4.5 Opus's "soul document." -- Show Notes: Everyone in Seattle hates AI https://www.geekwire.com/2025/viral-rant-on-why-everyone-in-seattle-hates-ai-strikes-a-nerve-sparks-debate-over-citys-tech-vibe/   Reranking partis...

It's an all AI research week, as we talk about the MIT Iceberg Index - a measure job exposure to AI on a per-task basis - an Anthropic study on the economic impacts of productivity improvements using a similar per-task measurement, and an Anthropic study which showed that models that learn to cheat on one thing can show bad behavior in other areas as well. -- Show Notes: MIT Iceberg Index - AI's E...

This week, we discuss the leaked draft executive order focused on federal pre-emption of state AI regulation and its baked-in assumption that LLMs are magic truth machines, the recent Cloudflare outage and the CEO's hand in putting together the detailed postmortem blog, and the trend in venture capital to fund things that are less adventurous. -- Show Notes: Exec order to stop state AI regulation...

We start this episode with a shoutout to the closure of Teen Vogue. In the main episode, we discuss Anthropic's recently released open-source evaluation of an LLM's "political even-handedness," a recent opinion article from a founder of an AI chatbot platform and whether the problem is the smut, and Marc Andreessen deleting his quote-tweet of the pope and the techno-optimist viewpoint overall. --...

In this week's episode, we talk about Amazon's recent lawsuit against Perplexity for them masking user agents while browsing the Amazon site, OpenAI's report on mental health-related conversations with ChatGPT, Character ai's recent decision to prevent minors from using their chatbots entirely, a quick OpenAI speed round, Andrew Cuomo's AI mpreg Schoolhouse Rock ad, and Zohran Mamdani engaging wit...

This week, we discuss Trump Media's Truth Social platform adding prediction markets, Elon Musk's Wikipedia-clone ego project, and an example of an organization standing on its principles and refusing federal funding. -- Show Notes: Truth Social is adding prediction markets https://www.bloomberg.com/news/articles/2025-10-28/trump-media-plans-to-enter-booming-prediction-markets-business Grokipedia h...

In this episode, we discuss Twitter beef between Jack Clark and David Sacks prompting an official Anthropic response, Twitch and the future of livestreaming, and how plausible it is that Elon Musk will have the robot army he claimed on Tesla's quarterly earning's call. Anthropic x Trump Admin https://importai.substack.com/p/import-ai-431-technological-optimism   https://x.com/davidsacks/status/197...

In this episode, we take a moment to run through updates on OpenAI topics from our last episode, before launching into two different economic policy topics - policy proposals Anthropic has gathered to address AI's economic fallout, and the debates on how to fix incentives in public markets. Show notes: OpenAI segment https://www.theverge.com/news/799938/japan-government-openai-sora?utm_source=subs...

This week we get into all the news from OpenAI, including third-party apps in ChatGPT, developments in the Sora AI video app's copyright policies, and the announced deal with AMD and how it's part of a pattern of circular funding deals. In the last third of the podcast, we talk about Cory Doctorow's Enshittification principle and how Amazon developed into what it is now. Show Notes: Apps in ChatGP...