Latest articles
3,574 articlesSort
19:08ENClaude Fable 5.1 and Mythos 5.1: Pricing, Safeguards and Hands-On TestsMatthew Berman · Forward FutureMatthew Berman examines Anthropic’s cost claims and benchmarks, an alternative cost estimate, and his own coding and design tests.
2:03:29ENLuke Bryan on Family, Football and Finding Humor in Country LifeGuest Luke Bryan · Host Theo VonThe country singer joins Theo Von for a wide-ranging conversation about growing up in the South, parenting and music.
3:43:42ZH → EnglishShougong Geng: The Dadaist in the Village Hardware ShopGuest 手工耿 · Host 罗永浩Luo Yonghao sits down with Shougong Geng for a serious talk about the road from stainless-steel slippers nobody would buy to being treated worldwide as a comic inventor—while Geng himself has always believed those things were useful.
20:47ENWhy OpenAI Plans to End Its Direct Model Agreement With CursorMatthew Berman · Forward FutureAfter SpaceX acquired Cursor, OpenAI proposed winding down its model-supply contract. Matthew Berman examines the dispute, competing providers and the risks of relying on another company’s models.
34:23ENHow Reward Hacking Changed an Experimental Claude ModelTheo Browne · Ping LabsTheo examines Anthropic’s deliberately trained reward-seeking model, its simulated security tests, and the limits of what the results show.
52:27ENAfter Two Dozen Police Departments Said No, They Built an Anti-Network-Effect AI CompanyA Wisconsin county had 300 gigabytes sitting on a cold case—and Peregrine's founders still refuse to take the credit.
1:35:29ENUber Burned a Year of Tokens in 13 Weeks. Retrieval Is How Agents WinHost Nathan LabenzMongoDB Field CTO of AI Pete Johnson argues that agent performance now hinges on retrieval quality, not on stuffing million-token context windows.
0:24ENTheo’s Interpretation of OpenAI’s Decision on Cursor and AstraTheoTheo connects the end of the partnership to concerns about model distillation and access to Astra.
1:05:56ENAaron Katz on ClickHouse, AI Revenue Durability and Enterprise DataGuest Aaron Katz · Host Harry StebbingsClickHouse’s CEO discusses agent workloads, open models, fundraising and the metrics he uses to judge durable growth.
15:20ENWhy Claude Code’s Limit Increase Means Less Usage for Current SubscribersTheo Browne · Ping LabsTheo explains the difference between Anthropic’s permanent baseline increase and the temporary allowance users were receiving, and criticizes how the change was communicated.
27:05ENOpenAI and Cursor’s Split: What Changes for DevelopersTheo Browne · Ping LabsTheo explains the proposed end of direct OpenAI model access in Cursor, the dispute over usage figures, and alternatives for developers.
1:29:22ENEno Reyes on AI Model Economics, Enterprise Control and Factory's StrategyGuest Eno Reyes · Host Harry StebbingsFactory's CTO discusses cost per outcome, model routing, open models, enterprise hiring and his expectations for the AI market.
25:27ENThe New Chips Challenging NVIDIA’s AI AdvantageTheo Browne, Jim Cramer, Jensen Huang · Ping LabsTheo compares local inference hardware with Chinese accelerators and OpenAI’s Jalapeño, focusing on memory, software and power efficiency.
18:57ENGLM-5.3-Flash: Ox Alpha Benchmarks, Costs and Coding TestsMatthew Berman · Forward FutureMatthew Berman reviews Z.ai’s open-weight model, compares benchmark costs and tries coding and presentation tasks against a competing model.
0:41ENCould Faster AI Agents Shift More Work to the Cloud?Matthew Berman · Forward FutureMatthew Berman reflects on a conversation about faster inference and local-computer limits. Whether the CPU is a bottleneck depends on the particular agent workload and execution environment.
2:10:42ENThe Exam Papers Nobody AuditsHost Nathan LabenzFrontier labs buy reinforcement-learning environments from a cottage industry of small vendors, and almost nobody checks whether those tests match the real work.
0:59ENA Laser Device Aims to Target MosquitoesMatthew Berman · Forward FutureMatthew Berman describes a proposed laser-based mosquito defense and the capabilities claimed in a short promotional video.
1:51ENA Vibe-Coding Skit About Authorization and SuperblocksMatthew Berman · Forward FutureA fictional customer-portal mishap illustrates server-side authorization risks before introducing Superblocks as a proposed governance tool.
17:15ENThe Hugging Face Incident: AI Evaluation, Sandbox Escape and Reward HackingMatthew Berman · Forward FutureMatthew Berman explains the reported OpenAI evaluation incident and the security and incentive failures it exposed.
1:26:43EN20VC: AI Funding, Model Competition and the Economics of AgentsGuest Jason Lemkin, Rory O’Driscoll · Host Harry StebbingsHarry Stebbings, Jason Lemkin and Rory O’Driscoll debate capital intensity, AI company economics and where agents may create value.
34:51ENGoogle Paid $10 Million for a Bankrupt Airline’s Data. That’s the Moat.Host Sarah Guo, Elad GilEon co-founders Ofir Ehrlich and Gonen Stein argue that models and compute are rented commodities — the only asset enterprises can actually keep is their data.
43:14ENWhy Theo Uses GLM-5.3-Flash for Coding Agent WorkTheo Browne · Ping LabsThe anonymous Ox Alpha model became GLM-5.3-Flash. Theo tests its pull-request triage, cost, multimodal features and limitations.
2:28:21ENLaura Delano on Psychiatric Treatment, Withdrawal and Informed ChoiceGuest Laura Delano · Host Theo VonThe author and advocate discusses her personal treatment history and the need for support when patients consider changing medication.
1:16:29ENGrant Sanderson and Alok Puranik: Positional Encodings Through Group TheoryGrant Sanderson, Alok Puranik · 3Blue1BrownStarting with assumptions about attention, they derive matrix exponentials and connect rotations, decay and Jordan blocks to RoPE and ALiBi.
Load moreShowing 24 of 3,574









