Subscribe
Share
Share
Embed
Daily AI news and research, distilled. UpNext AI breaks down the most important developments in artificial intelligence—from major industry moves to cutting-edge papers.
Welcome to the UpNext AI podcast. It's Monday, May 18th, 2026, and here's what matters in AI today.\n\nFirst up, OpenAI is expanding ChatGPT into personal finance. TechCrunch reports the company has launched a preview of personal finance tools for ChatGPT Pro subscribers in the U.S., letting users connect bank accounts and other financial accounts through Plaid. Once those accounts are linked, ChatGPT can show a dashboard with portfolio performance, spending, subscriptions, and upcoming payments.\n\nOpenAI says people already ask ChatGPT financial questions at massive scale, and this product is clearly an attempt to turn that behavior into a dedicated workflow instead of a generic chat. The company also says the feature is coming first to the web and iOS for Pro users, with a staged rollout that references May 27, and that synced data is removed within 30 days after a user disconnects a service.\n\nThe practical significance here is pretty straightforward. This is OpenAI moving from advice about money to software that actually sits closer to the financial data itself. That does not make it a bank, and it does not tell us yet how broadly this will roll out, but it does show where consumer AI products are heading: less chatbot, more control panel for high-stakes daily life.\n\nFor the second act today, a funding story that says something about where applied AI money is still going. TechCrunch reports that Nectar Social has raised a 30 million dollar Series A led by Menlo Ventures and its Anthology Fund, which was created alongside Anthropic.\n\nNectar Social describes itself as an agentic operating system for marketers. According to TechCrunch, its AI agents help brands handle social activity, moderation, creator workflows, competitive intelligence, and commerce conversations end to end. The company also says it has data partnerships with Meta and Reddit so customers can pull information from multiple platforms into one place instead of juggling separate tools.\n\nThe bigger point is that even in a market flooded with AI copilots and agents, investors are still backing vertical systems that plug directly into revenue functions. In this case, that’s marketing and social commerce. Nectar says clients include Liquid Death, Figma, and e.l.f. Beauty, and that the new capital will go toward hiring across applied AI, engineering, and go to market. So while frontier model news still gets the spotlight, this is a useful reminder that a lot of the commercial race is now about packaging AI into one workflow a business will actually pay for.\n\nNow to the research section, and today that’s really an operations story dressed as a launch. Giotto.ai says it is launching portable AI for enterprises, with advanced reasoning that can move from cloud to workstation. In the company’s press-release style announcement, the pitch is a general-purpose model and operating system that organizations can run on their own GPUs, through Giotto-managed capacity, or via pre-installed appliances, with configurations starting from a single GPU.\n\nIn plain English, the idea is simple: instead of assuming advanced AI has to live in a remote cloud run by a hyperscaler, Giotto is arguing that companies and governments will want reasoning systems they can install, govern, and operate inside their own infrastructure.\n\nThat matters because the tradeoff in enterprise AI is no longer just model quality. It is also latency, control, compliance, sovereignty, and whether a company wants a core system to depend on an outside API. Giotto says that independence is the main feature here, especially for sectors like public administration, defense, finance, healthcare, energy, manufacturing, and telecom.\n\nWhat we do not have in the provided reporting are the hard details you’d want to fully evaluate the claim. There are no benchmark numbers, no hardware specifics beyond the single-GPU framing, and no concrete deployment case studies in the announcement we have. So the takeaway is less that Giotto has definitively changed the market today, and more that the industry direction is getting clearer: advanced reasoning is increasingly being packaged to run closer to where the work and the data already are.\n\nBottom line: for enterprise buyers, the next meaningful AI differentiator may be not just how smart a system is, but how much of it you can actually control.\n\n...Are you building apps with voice? Elevate your app's voice capabilities with ElevenLabs. Their API is a game changer for embedding dynamic, responsive voice interactions in your applications, providing unprecedented realism, flexibility and latency. In fact, you're listening to one of their voices - right - now. If you are a developer looking to elevate user experience with natural voice interfaces, this is your solution. Visit up next dot fm slash eleven to check out their latest offerings. ...\n\nMicrosoft is retiring Teams’ Together Mode. The Verge reports that the feature, introduced during the pandemic, used AI to cut out a participant’s head and shoulders and place them into a shared virtual meeting space. Microsoft is phasing it out as part of a simpler Teams experience.\n\nIn the UK, Simon Willison highlights a new intervention from the Government Digital Service in the debate over the NHS moving away from open source. The key recommendation from GDS: keep open by default, with closure used sparingly and deliberately.\n\nAnd Simon Willison also notes a new release called datasette-llm-limits 0.1a0. The plugin works with datasette-llm and datasette-llm-accountant to set per-user, or global, spending limits for LLM usage inside Datasette. Small tool, practical idea.\n\nBefore we wrap up, a quick note: this podcast is generated with the assistance of AI and is intended for informational purposes only. All referenced articles, research, and commentary remain the property of their original authors and publishers.\n\nIf you enjoyed this episode, don't forget to subscribe, rate, and leave us a review! And that's your briefing for today. Full source links are in the episode notes, and we'll be back tomorrow with what's up next!