Vol. I · No. 288
“All the News That Fits the Feed”
Est. 2026
Weather: A clear Thursday
in the Levantine morning

The Daily Fledglingהגוזל היומי



Morning Edition · Wednesday, October 15, 2026 · Price: One Click


Edited & compiled by Nadav Solomon

Developer Tools

Claude Code Desktop Learns to Juggle Sessions

Nested agents, a live to-do rail, /btw side tabs—and a v2-feel redesign for parallel work

Pick an environment, spin sessions, watch the sidebar track what each Claude is doing.בוחרים סביבה, מפעילים סשנים, והסיידבר עוקב אחרי מה שכל Claude עושה.Image: תמונה: @ClaudeDevs on X

Anthropic’s ClaudeDevs team shipped a sweeping Claude Code desktop update aimed at one pain: running lots of agent sessions side by side without losing the plot. Parallel tasks now spawn nested sessions that report back into a readable main thread. A live to-do list tracks Claude’s plan; a recap card appears when it needs you; the side rail shows where it runs, its context and its diffs.

You can save folders, SSH hosts or cloud setups as named environments (with optional worktrees), restart from recent folders in one click, ask side questions with /btw in their own tab, and one-click spin follow-ups Claude suggests mid-run. Thariq (trq212) called it almost a v2—Claude now “knows” it can spawn subsessions. Update the desktop app and start a new session to try it.

Sources: x.com/ClaudeDevs/status/2110445844675838148 · x.com/trq212/status/2110463962877989276

Voice AI

Grok Voice 3.0 Takes the Speech Crown

#1 on Artificial Analysis at 86.9—and SpaceX says it already closes 70% of Starlink support calls

SpaceX Super Intelligence (xAI’s voice arm on the feed) launched Grok Voice 3.0: a speech-to-speech model that answers in under a second, posts at $0.08/min in the API and Voice Agent Builder, and—per SpaceX—resolves about 70% of Starlink support calls on its own with ~40% lower latency than prior frontier voice agents. Artificial Analysis put it at #1 on the Speech to Speech Index (86.9), ahead of OpenAI’s GPT-Live-1 Astra (83.2) and Gemini 3.8 Live Extended Thinking (82.6).

Sources: x.com/SpaceXSI/status/2110416580899168305 · x.com/ArtificialAnlys/status/2110421964095713743

Artificial Analysis table: Grok Voice 3.0 leads speech-to-speech metrics vs GPT-Live-1 and Gemini 3.8 Live
The scoreboard: Grok Voice 3.0 leads the AA speech-to-speech table across overall, reasoning and task metrics.לוח התוצאות: Grok Voice 3.0 מוביל בטבלת speech-to-speech של AA בכל המדדים העיקריים.Image: תמונה: @SpaceXSI on X

Science Desk

Argon Clocked In at the Lab Bench

Enzymes in hours, a 60-year plasma conjecture challenged, and 10,000 scientists get free Ultra Antigravity

A day after the global Gemini 4 Argon launch, DeepMind posted the research receipts: zero-shot enzyme engineering that designed five lab-verified serine esterases in hours; novel 3D magnetic-field designs with Maryland collaborators that further counter Grad’s nearly 60-year-old plasma-physics conjecture; and speedups for quantum error-correction decoders turned into optimized hardware megakernels. To push more of that work, DeepMind is giving 10,000 scientists six months of free highest-tier Antigravity access. Sundar Pichai amplified a Fourier-epicycle pelican bike clip as “creativity in the prompting and the model.” Google AI demos include an interactive photosynthesis atlas Argon built in Antigravity.

Sources: x.com/GoogleDeepMind/status/2110421861435879476 · x.com/sundarpichai/status/2110443819477156150 · x.com/GoogleAI/status/2110392574754107405

Textbook biology, browsable: Argon builds a layered photosynthesis atlas in Antigravity.ביולוגיה מספר לימוד, בגלישה: Argon בונה אטלס פוטוסינתזה בשכבות ב־Antigravity.Image: תמונה: @GoogleAI on X
Collaboration

ChatGPT Space: Shared Docs That Stay Alive

DevDay’s team room—turn a messy chat into a living plan your coworker can edit with the model

ChatGPT’s official account highlighted ChatGPT Space, introduced at DevDay: a shared workspace where a team co-writes living documents with ChatGPT. The pitch is simple—messy conversation becomes an organized plan, then a teammate jumps in to edit alongside the model. It is the consumer/team counterpart to yesterday’s model launches: collaboration furniture, not another benchmark chart.

Read the post: x.com/ChatGPT/status/2110424821331034550


Open Agents

Hermes Agent Crosses the Threshold

v0.22.0 ships—and 50+ connectors go live with no account required

Nous Research tagged Hermes Agent v0.22.0 “The Threshold Release,” then followed with a connectors launch: 50+ connectors available to everyone without an account, optional Nous portal sync for settings, multi-account aliases, and permission limits (read/write/destructive). Street chatter—and a few earnest WhatsApp lawyer testimonials—suggest Hermes is having a moment as a personal agent people actually install.

Sources: x.com/NousResearch/status/2110385143643091204 · x.com/NousResearch/status/2110470076595835353

Briefly Noted

Singapore, Design, Launchers & Robot Arms

Anthropic’s fifth APAC office, Mistral buys New Material, Epic vibe-codes a launcher, and AA grades robot models

Anthropic opened a Singapore office—its fifth in Asia-Pacific—and signed an MOU with Singapore’s government on AI safety and responsible adoption. Mistral welcomed design studio New Material in-house to “bring design to mission-critical AI.” Roundtable says Epic Games vibe-coded a new launcher that beats Steam on cold start (5.47s vs 10.07s) with native Linux. Concat landed as a free CapCut-style open-source editor (no watermark, local captions/TTS). Artificial Analysis launched AA-Robotics v0.1; GPT-6 Astra leads (69) ahead of GPT-6.1 Sol (65) and Gemini 4 Argon (55). The EU AI Office sent information requests to OpenAI and Anthropic over testing incidents. Anthropic published an official Claude Code “code-modernization” plugin for COBOL/legacy Java/C++/.NET. Claude Platform OpenTelemetry tracing hit public beta. Qwen cheered TokenSpeed numbers: Qwen3.8-Flash-Next at 685 output tok/s per user on 4× GB300. OpenAI’s EU economic note: more than half of EU enterprise AI use now involves agents; small-business users up 8×.

Sources: x.com/AnthropicAI/… · x.com/MistralAI/… · x.com/ArtificialAnlys/… · x.com/ClaudeDevs/… · x.com/RoundtableSpace/…

On the Air · Radio

The Fledgling Wireless

Wren & Robin read you this morning's dispatches, with banter.

Edited & published by Nadav Solomon
בשידור · רדיו

האלחוט של הגוזל

רֶן ורובין מגישים את ידיעות הבוקר, עם קצת צחוקים.

בעריכת והוצאת נדב סולומון
Published by Nadav Solomon · Printed on the Fledgling Press · Set in Georgia & Times