AI News Today | Julian Goldie Podcast

Sakana: BETTER Than Fable 5?

14 min · 25. juni 2026
episode Sakana: BETTER Than Fable 5? cover

Beskrivelse

Sakana Fugu Ultra vs Fusion vs Claude Opus 4.8 (42 Builds + Goldie Bench): Is It Fable 5-Level?The video reviews Sakana Fugu Ultra after several days of release, testing whether it matches “Fable 5-level intelligence” claims by running 42 builds using the same prompts across Fugu Ultra, Fusion (OpenRouter), and Claude Opus 4.8, plus Goldie Bench leaderboard comparisons. The script explains Fugu as a multi-agent orchestration system that fuses outputs from closed and open models, contrasting it with Fusion’s pay-per-token API pricing versus Fugu’s flat subscription with token limits and frequent lockouts, slow responses, and one-shot workflows. Side-by-side build demos (solar system, runner, fireworks, dungeon crawler, open world) show mixed results but generally stronger outputs from Fugu Ultra, with Fusion often close or better on some examples and Opus 4.8 frequently failing. Benchmarks cited include Fugu leading LiveCode Bench but trailing Fable 5 on SWE Bench Pro and Terminal Bench, and a brief comparison of Fugu Ultra vs cheaper Fugu Mini. The episode ends by promoting the AI Profit Boardroom agent operating system integrating Sakana, Fusion, Claude, and other tools.00:00 [https://www.youtube.com/watch?v=uhcWUS4f0vc] Fugu Ultra Overview00:45 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=45s] Testing Setup and Limits01:26 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=86s] Models and Method02:02 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=122s] What Is Fugu03:00 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=180s] Leaderboard and Pricing04:12 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=252s] Build Demos Part 106:21 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=381s] Build Demos Part 208:33 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=513s] Open World Comparison09:23 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=563s] Benchmark Breakdown10:16 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=616s] Fugu Mini vs Ultra12:03 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=723s] Final Verdict12:13 [https://www.youtube.com/watch?v=uhcWUS4f0vc&t=733s] Boardroom and Outro

Kommentarer

0

Vær den første til at kommentere

Tilmeld dig nu og bliv en del af AI News Today | Julian Goldie Podcast-fællesskabet!

Kom i gang

1 måned kun 9 kr.

Derefter 99 kr. / måned · Opsig når som helst.

  • Podcasts kun på Podimo
  • 20 lydbogstimer pr. måned
  • Gratis podcasts

Alle episoder

579 episoder

episode Hermes Apollo, Astros & Oracle are INSANE! cover

Hermes Apollo, Astros & Oracle are INSANE!

Hermes Apollo, Astros & Oracle: The 24/7 AI Command Crew for Real-Time Trends, Content & MemoryThe script introduces three new Hermes-based systems—Hermes Apollo, Hermes Astros, and Hermes Oracle—designed to run 24/7 as a “command crew” for real-time voice control, trend monitoring, and content creation. Oracle pulls the latest trending topics, ranks them, opens them on X, helps draft social posts, and can publish SEO content directly to WordPress. Apollo is a voice-activated agent with a wake word, real-time responsiveness via the ChatGPT API, computer-use actions (e.g., opening websites), and conversation/build history. Astros tracks niche keywords and competitors in real time, generates titles, angles, and hooks, and feeds workflows like video, SEO, and notebook outputs. All tools write into a shared Obsidian “Memory Galaxy” that updates continuously, supports daily/weekly briefings, and improves over time, with token-minimization playbooks and free model APIs to reduce costs, available via the AI Profit Ballroom community.00:00 [https://www.youtube.com/watch?v=W48NhfblaIA] Hermes Systems Overview00:12 [https://www.youtube.com/watch?v=W48NhfblaIA&t=12s] Oracle Trending to Posts00:42 [https://www.youtube.com/watch?v=W48NhfblaIA&t=42s] Command Crew Explained00:57 [https://www.youtube.com/watch?v=W48NhfblaIA&t=57s] Apollo Voice Jarvis02:05 [https://www.youtube.com/watch?v=W48NhfblaIA&t=125s] Astros Trend Engine03:41 [https://www.youtube.com/watch?v=W48NhfblaIA&t=221s] Memory Galaxy Brain04:39 [https://www.youtube.com/watch?v=W48NhfblaIA&t=279s] Use Cases and Leads05:16 [https://www.youtube.com/watch?v=W48NhfblaIA&t=316s] Weekly Daily Briefings07:08 [https://www.youtube.com/watch?v=W48NhfblaIA&t=428s] Always On Workflow Loop08:29 [https://www.youtube.com/watch?v=W48NhfblaIA&t=509s] Any Niche Watchlists09:20 [https://www.youtube.com/watch?v=W48NhfblaIA&t=560s] Token Cost Minimization10:04 [https://www.youtube.com/watch?v=W48NhfblaIA&t=604s] Join AI Profit Ballroom10:59 [https://www.youtube.com/watch?v=W48NhfblaIA&t=659s] Testimonials and Setup Ease11:50 [https://www.youtube.com/watch?v=W48NhfblaIA&t=710s] Final Call to Action

9. juli 202611 min
episode GPT 5.6 is out TOMORROW! cover

GPT 5.6 is out TOMORROW!

GPT 5.6 Launch Prep: 4 Moves to Be Ready on Day One (Sol, Terra, Luna)The script explains how to prepare for OpenAI’s GPT 5.6 public launch on Thursday, which includes three models—Sol (flagship), Terra (lower cost), and Luna (fastest and cheapest)—with preview access expanding globally. It outlines four prep moves: log every daily task, time-audit tasks by duration and frequency, minimize tokens before scaling (using free/local models and token “playbooks” like Caveman, Ponytail, RTK, Headroom, and techniques such as shorter paid replies and matching reasoning effort to the task), and build an agentic operating system around the audit so GPT 5.6 can plug into existing automations immediately. The script argues the system around the model matters more than the model itself and promotes the AI Profit Bot as a ready-made Agent OS with agents, memory, free APIs, tutorials, community support, and coaching calls.⁠00:00⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc] GPT 5.6 Launch Hype⁠00:39⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=39s] Three Model Breakdown⁠01:12⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=72s] Four Prep Moves⁠03:21⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=201s] Audit Your Week⁠04:15⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=255s] Token Saving Playbooks⁠04:43⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=283s] Pick the Right Model⁠05:16⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=316s] Cut Costs and Effort⁠07:21⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=441s] Systems Beat Models⁠08:27⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=507s] AI Profit Bot Offer⁠09:45⁠ [https://www.youtube.com/watch?v=3TD2jmHoUyc&t=585s] Wrap Up

I går9 min
episode Claude Loops: Build Agents That Run Without You cover

Claude Loops: Build Agents That Run Without You

5 Claude Agent Loops That Run Autonomously (Build, Plan, Ship, Watch News & Write SEO) | Agent OSThe script explains five autonomous “Claude agent loops” built inside an Agent OS that reduce manual prompting by using a repeatable cycle where an agent acts, an independent judge checks the work against a defined goal, and the loop repeats until it passes, optionally with a human approval step. The five loops include: a loop-engineering builder/judge setup to build deliverables to a spec while optimizing cost with different models; a Kanban workflow with planner, builder, and reviewer roles that moves tasks through triage, running, blocked, and done; a pipeline that takes ideas from inbox to shipped with one-time plan approval and saved outputs in a vault; an “Oracle” timed news loop that finds and ranks AI news and produces content hooks while logging to Obsidian memory; and an SEO pipeline that generates keywords and one-click SEO articles. A shared memory system updates automatically to compound performance, and the video promotes access via the AI Profit Boardroom with coaching and community support.00:00 [https://www.youtube.com/watch?v=zRir38lm8eQ] Autonomous Agent Loops02:26 [https://www.youtube.com/watch?v=zRir38lm8eQ&t=146s] Old Way vs New Way03:23 [https://www.youtube.com/watch?v=zRir38lm8eQ&t=203s] Loop Engineering Builder04:22 [https://www.youtube.com/watch?v=zRir38lm8eQ&t=262s] Kanban Multi Agent Board05:25 [https://www.youtube.com/watch?v=zRir38lm8eQ&t=325s] Idea to Shipped Pipeline07:20 [https://www.youtube.com/watch?v=zRir38lm8eQ&t=440s] Oracle News Watcher08:42 [https://www.youtube.com/watch?v=zRir38lm8eQ&t=522s] SEO Content Pipeline09:10 [https://www.youtube.com/watch?v=zRir38lm8eQ&t=550s] Shared Memory System09:55 [https://www.youtube.com/watch?v=zRir38lm8eQ&t=595s] Wrap Up and Access

I går11 min
episode Codex + HY3 is Insane (FREE!) cover

Codex + HY3 is Insane (FREE!)

Use Codex for FREE with Tencent HY3 (OpenRouter API) + OmniRoot BackupThe video shows how to use Codex for free by plugging Tencent’s new open-source HY3 model into Codex via a free OpenRouter API (available for about two weeks, until around July 21). The creator demonstrates a custom setup with goal mode and a workspace where you can preview and manage past builds, plus a separate “HY3 Coder” that runs directly on HY3 without Codex. They explain the workflow: Codex acts as the hands (writes files, runs commands) while HY3 is the brain (agentic coding and tool calling), and you can swap the model in the config or have Claude/Claude Desktop set it up. If rate-limited, you can switch to OmniRoot, another free API, and also use HY3 inside Hermes as a free agent profile.00:00 [https://www.youtube.com/watch?v=eFx4AmwiQtc] Free Codex With HY300:35 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=35s] Workspace Demo01:08 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=68s] Setup Via OpenRouter02:10 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=130s] HY3 Coder Alternative02:44 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=164s] Hands And Brain Explained03:36 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=216s] Hermes And OmniRoot04:29 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=269s] Live Task Switching05:23 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=323s] HY3 Performance Benchmarks06:11 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=371s] Token Saving Workflow07:16 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=436s] Get The Full Agent OS08:22 [https://www.youtube.com/watch?v=eFx4AmwiQtc&t=502s] Final Thanks

I går8 min