Short summary pods for busy people.
Subscribe, combine, create.

Subscribe to any of the podcasts in the library using the RSS feed (no account needed!)

Combine many podcasts into your own custom podcast to get them all to a single feed

Create a custom summary pod with your own sources

Featured Pods

anthropicopenaideepmindblognewssimonwillison@karpathy@sama@AnthropicAI@swyx@levelsio@yoheinakajima@dair_ai@rowancheung@rasbt@bcherny@steipete

Add this pod to your own podcast

All episodes from this podcast will be fed into your own.

Jul 8, 2026Here is today's AI Daily for Thursday July 9th. Yesterday’s biggest model story was xAI’s public launch of Grok 4.5, positioned less as the absolute benchmark king and more as a fast, cost-efficient coding and agent model. xAI says it was trained specifically for coding and agents, with Cursor as a training and distribution partner. The headline numbers circulating from Artificial Analysis are a 500,000-token context window, pricing at $2 per million input tokens and $6 per million output tokens, and near-frontier rankings on broad intelligence and coding-agent evals. The takeaway: xAI is now a credible player in the coding-agent market, especially for teams optimizing dollars per completed task. OpenAI also had a busy day yesterday. GPT-Live launched in ChatGPT as a next-generation full-duplex voice system that can keep a conversation flowing while delegating harder work, like search or deeper reasoning, to a frontier model in the background. OpenAI also published work on coding evaluations, saying it audited SWE-Bench Pro and found about 30% of tasks broken, retracting its prior recommendation to treat it as a leading coding benchmark. And OpenAI says GPT-5.6 Sol launches today, so the competitive pressure will only intensify. A third story is Jarred Sumner’s detailed writeup on rewriting Bun from Zig to Rust with agentic engineering. Simon Willison highlighted it as a major case study: a huge conformance test suite, parallel agent workflows, adversarial review, and process-level fixes helped make a million-line rewrite plausible. The estimated pre-merge token bill was about $165,000 at API pricing, and the Rust port has reportedly already been live inside Claude Code for weeks. Finally, coding-agent competition is broadening beyond the big labs. Cognition launched SWE-1.7, built on a Kimi base, claiming near-frontier coding performance at very high speed, while NVIDIA and LangChain promoted an open enterprise agent stack with much lower inference costs. The trend is clear: the frontier is shifting from “best model” to “best system.” Speed, price, harness design, eval quality, and real workflow integration now matter as much as raw intelligence. Thank you for listening to AI Daily from The Daily FM. See you tomorrow!

World Cup Daily

fifafoxsportsbbcussoccer

Add this pod to your own podcast

All episodes from this podcast will be fed into your own.

Jul 9, 2026Here is today's World Cup Daily for Thursday July 9th. The quarter-final spotlight is sharpening, and England against Norway has become the glamour tie. The BBC is framing it around Harry Kane versus Erling Haaland: two elite strikers, one semi-final place, and a Norway side riding the high of eliminating Brazil. England, meanwhile, were given an offsite family day yesterday after the Mexico win, but there are concerns. Jordan Henderson has had surgery on a broken arm, though he has not been ruled out for the rest of the World Cup, and four England players are walking a yellow-card tightrope before a possible semi-final. Argentina’s dramatic comeback over Egypt is still producing fallout. Yesterday, the Egyptian FA lodged a complaint with FIFA, alleging double standards in officiating and asking for officials to be removed from the tournament. The wider question now is whether Argentina and Lionel Messi are being treated favorably, especially after Messi admitted he felt he had “let everyone down” with his missed penalty before helping rescue the match. This is becoming one of the tournament’s biggest integrity debates. For the United States men’s national team, the postmortem continues after the loss to Belgium. Mauricio Pochettino said the Americans “weren’t good enough” and insisted the uncertainty around Folarin Balogun’s suspended ban did not decide the match. That matters, because the U.S. exit is now being viewed less as one bad night and more as a reminder of the gap between a promising host-nation run and the ruthlessness required at the last-eight level. There is also coaching churn after the knockouts. Portugal confirmed Roberto Martinez’s departure, Mexico moved on from Javier Aguirre and appointed Rafael Marquez, and Croatia’s Zlatko Dalic stepped down after nine years. The trend is clear: this expanded World Cup has created more opportunity, but also harsher judgment. Big nations are not just expected to qualify or entertain — they are expected to survive chaos, manage discipline, and reach the business end. Thank you for listening to World Cup Daily from The Daily FM. See you tomorrow!

Financial Markets

bloombergcnbcfinanceft@WSJmarkets@markets@business@CNBCnow@ReutersBiz

Add this pod to your own podcast

All episodes from this podcast will be fed into your own.

Jul 8, 2026Here is today's Financial Markets for Thursday July 9th. Renewed Middle East fighting is back at the center of markets. Yesterday, President Trump said an interim deal aimed at ending the Iran conflict was “over,” and Yahoo Finance says U.S. stock futures slipped after fresh U.S. strikes. Reuters reported the S&P 500 ended lower, while oil climbed on supply-disruption fears. Bloomberg also said traffic through Hormuz has slowed sharply, raising the risk that energy prices become a bigger inflation problem again. Safe-haven and rates markets are sending a mixed message. The Wall Street Journal reported gold was roughly flat despite renewed fighting and Fed minutes, while Japanese government bonds fell, tracking weaker U.S. Treasurys. Bloomberg also noted Jupiter Asset Management cut U.S. Treasury holdings to zero in one major bond fund, shifting toward European government debt and emerging markets. That points to a broader concern: investors are not just reacting to headlines, they are rethinking where duration risk looks most attractive. In technology, the AI supply chain still has strong investor demand. Reuters and Bloomberg report SK Hynix’s U.S. share sale was more than seven times oversubscribed, as the South Korean memory-chip maker prepares to price the offering today. That is a notable contrast with recent worries about stretched AI valuations. Investors may be questioning some high-multiple tech stocks, but they still want exposure to companies seen as critical to AI infrastructure. Apple also made a major supply-chain move yesterday, announcing a 30 billion dollar deal with Broadcom for chips through 2031, including expanded U.S.-based manufacturing in Colorado. Broadcom helped lead gains among chip stocks even as broader indexes weakened. The big trend is a market being pulled in two directions. Geopolitics and oil are pushing investors toward caution, while AI infrastructure and domestic chip manufacturing continue to attract capital. If oil keeps rising, the Fed pause story could get harder to defend; if energy stabilizes, selective tech leadership may remain the market’s strongest support. Thank you for listening to Financial Markets from The Daily FM. See you tomorrow!

Latent Space in 3 minutes

api.substack.com

Add this pod to your own podcast

All episodes from this podcast will be fed into your own.

Jul 9, 2026Here is The Daily FM summary of the Latent Space that aired on Wednesday July 8th. The episode featured Modal CTO Akshat Bubna, alongside cofounder Vibhu, in a wide-ranging conversation about how AI infrastructure is changing as agents, inference, and bursty GPU workloads become central to modern software. Akshat began by revisiting Modal’s origin story. The company did not start as a GPU inference platform, but as an attempt to build a better runtime than Kubernetes for compute-heavy, bursty workloads. The key idea was that developers should not have to manage piles of YAML or Kubernetes configuration just to run jobs with custom images, accelerators, or fast scaling. Modal’s answer was “self-provisioning workloads”: infrastructure requirements live next to the code, often as decorators, so the system can spin up the right compute automatically. A major theme was the shift from developer experience to what Akshat called agent experience. Modal has apparently reorganized its SDK thinking around how AI agents use infrastructure. His argument was simple: if it is painful for humans to read and modify hundreds of Kubernetes files, it is also painful for agents. Typed, colocated infrastructure definitions are easier for coding agents to modify, test, and observe. The hosts pushed on whether this still matters if humans are not reading code as much anymore, and Akshat said observability may now be even more important: agents can change code, but humans still need dashboards, logs, and judgment to understand what happened. The conversation then moved into Modal’s current identity. Akshat described it as a cloud platform with primitives built from scratch for AI applications, covering inference, training, batch processing, and sandboxes. He emphasized that Modal is not trying to replace always-on web hosting. Its sweet spot is specialized compute that scales up and down fast, across GPUs, CPUs, regions, and unusual workloads. One of the most interesting sections covered sandboxes. Modal built sandbox APIs as early as 2023, before coding agents had fully taken off, and even used an early “small developer” agent loop as an example. Back then, models tended to diverge after a handful of iterations. In hindsight, the hosts joked, the winning move would have been to collect all those failures, build benchmarks and RL environments, and turn them into a billion-dollar agent company. Akshat also explained that Modal’s biggest use case today is elastic inference for custom models, especially in audio, video, robotics, and computational biology. Companies like Suno and Runway may train models elsewhere but use Modal to deploy and autoscale them. Modal has gone deep on cold starts, GPU snapshotting, and regional autoscaling, because production inference is not just “find a GPU and run a model”; it involves tail latency, reliability, request delivery, and fast scaling across regions. A technical highlight was speculative decoding. Akshat explained that a smaller draft model predicts tokens ahead, while the larger model verifies them. If many drafted tokens are accepted, inference can become two to four times faster without quality loss. Modal’s open-source DFlash work uses block-based speculation, and its new Auto Endpoints aim to make optimized open-model serving easier while still letting users eject into full code when they need customization. The episode also touched on distributed training, RDMA networking, private IPv6 overlay networks, and why Modal runs across 17 cloud and neocloud providers rather than owning data centers. Akshat framed Modal as a “supercloud” software layer, with its own reliability and scheduling layer on top of many providers. Near the end, the discussion broadened to AI infrastructure trends: agent sandboxes, CI for coding agents, continual learning, auto-research, robotics, drug discovery, and whether future video generation may be orchestrated by agents rather than single video models. The final takeaway was that Modal’s bet on developer experience has evolved naturally into agent experience, and that the infrastructure primitives that once seemed niche now look central to the next generation of AI products. Thank you for listening to Latent Space in 3 minutes from The Daily FM. See you tomorrow!

View the whole library →

Subscribe to one of these pods with the RSS feed, or create your own →

Short summary pods for busy people.Subscribe, combine, create.

Featured Pods

Short summary pods for busy people.
Subscribe, combine, create.