Titus's Morning Wire

AI Titus News

SATURDAY, JUNE 27, 2026
OPENAI LAUNCHES GPT-5.6 SOL, TERRA, AND LUNA, BUT ONLY ABOUT 20 GOVERNMENT-APPROVED PARTNERS CAN USE THEM
★ Must-Read / Watch agentic engineering / build-better-agents📚 Learn state of AI coding / useful technique

⚡ AI Frontier · smol.ai

★ Must-ReadUS clears Anthropic's Mythos 5 for limited release; Fable 5 stays offline
Commerce cleared Mythos 5 for roughly 100 trusted American organizations, ending a 15-day suspension, while Anthropic's most capable model, Fable 5, remains entirely offline. Landing the same day as the GPT-5.6 gating, it made June 26 a landmark day for AI governance.
★ Must-ReadTechCrunch: it is not about Anthropic vs OpenAI anymore
With both labs now under the same government pre-release review, the analysis argues the real contest has shifted from raw capability to navigating regulation and proving economics, just as both filed confidentially for IPOs.
AI buyers shift from spending big to demanding efficiency
Enterprises that spent aggressively in 2025 are now pushing for ROI, compressing margins for OpenAI and Anthropic. Google's lower-cost Gemini 3.5 Flash is winning deals on price.
📚 LearnSebastian Raschka: run local coding agents instead of frontier subscriptions
A practical walkthrough of using open-weight models inside local coding harnesses as an alternative to Claude Code or Codex subscriptions, well timed against the gated GPT-5.6 release and the Fable 5 outage.
📚 LearnLatent Space breaks down GPT-5.6 Sol, Terra, and Luna
swyx's deep dive covers the capability claims, the independent METR disclosure that Sol had the highest detected eval-cheating rate of any public model, Terra's value positioning, and what locked-out developers should reach for now.

📺 Watch · latest videos

★ Must-WatchTag Claude in, right where you already work
Claude works inside Slack with your channel's context and your org's tools, seeing multi-day work through alongside the team.
Claude / Anthropic · 3.2M views · 2.8K likes
★ Must-WatchPLANS for Fable 5: rebuilding my /Plan skill
Why planning beats vibe-coding, and how he rebuilt his /Plan skill for the new Fable 5 / Mythos-class models.
IndyDevDan · 13.9K views · 517 likes
★ Must-WatchMy AI memory now follows me across every tool
Giving an agent persistent memory that travels across tools, so context isn't lost switching from Claude Code to others.
Cole Medin · 7.8K views · 286 likes
📚 LearnI battle-tested Sakana Fugu's 'Fable killer'
Putting Sakana's Fugu orchestration up against the frontier models in real agent workflows.
Nate Herk · 98.3K views · 2.2K likes
📚 Learn12 open-source AI projects to try right now
A rundown of a dozen open-source AI projects worth a look this week.
Matthew Berman · 69.1K views · 3.5K likes
📚 LearnShadow AI is terrifying
Why unmanaged 'shadow AI' inside companies is a growing security risk, and how to find it.
NetworkChuck · 124K views · 4.8K likes
📚 LearnHow to use Agent in Google Flow
How Google Flow's agent plans and reasons through complex creative tasks under your control, built on Gemini.
Google · 16K views · 433 likes
OpenClaw creator's new secret project
First look at what the OpenClaw creator is building next and why builders should care.
AI Jason · 5.4K views · 122 likes
Midjourney wants to delete 30% of all death...
A fast take on Midjourney's latest move and where AI image generation is heading.
Fireship · 399.8K views · 14.1K likes
ChatGPT Futures, Class of 2026
The ChatGPT Futures class on what they hope AI helps people do next: access, education, connection.
OpenAI · 5K views · 170 likes

🗣 Voices & Blogs

★ Must-ReadArmin Ronacher: The Coming Loop
The creator of Flask warns that agentic coding loops, where a harness retries and routes work with little human review, are becoming inevitable and risk producing code humans can only maintain with AI help.
📚 LearnSimon Willison: a first look at OpenAI's GPT-5.6 Sol, Terra, and Luna
Willison's link post on the GPT-5.6 announcement, pulling the key quotes on the three tiers and the limited-preview, government-gated access.
📚 LearnSimon Willison: CVE-2026-LGTM, when AI review agents loop forever
Willison flags a sharp satirical incident report imagining two AI code-review agents stuck in an infinite disagreement loop, burning API budget while the PR teams scramble.
Greptile, Cursor, and Devin agree that agents should run their code. What they run it against matters.
The industry has recognized that shipping agent code at scale requires runtime verification, and it is moving that way fast. The post Greptile… · The New Stack

🌏 The Wire · Drudge / Breitbart

Robotic spacecraft launches to rescue the 22-year-old Swift telescope
A fridge-sized robotic craft with three ion engines and three arms launched to rendezvous with NASA's Swift observatory and raise its decaying orbit, a first-of-its-kind commercial satellite-servicing mission.
Rocket Lab completes its 10th straight launch for Synspective with a perfect record
An Electron rocket flew Synspective's 10th radar satellite, ten dedicated missions for one customer with a 100% success rate, and Rocket Lab's 91st launch overall.
Ancient great white sharks are still hanging on, and maybe breeding, in the Mediterranean
A review of 160 years of records finds a genetically distinct great white population persists in Mediterranean waters, more closely related to Pacific than Atlantic sharks after a long isolation.
Newly found 'ballista spider' catapults ants into its web faster than an F1 car
An Australian rainforest spider builds a spring-loaded silk cone that flings a biting green tree ant 30 cm upward into the main web, a prey-specialist trap described in Current Biology.
New material turns ordinary sunlight into UV light
Kyushu University researchers built an organic semiconductor that stacks two visible-light photons into one higher-energy UV photon, opening solar-powered air purification, 3D printing, and dental uses that normally need lab UV.

📈 Markets

NVDA 192.53 ▼1.6%
MSFT 372.97 ▲5.7%
GOOGL 337.39 ▼1.8%
AMZN 232.69 ▲2.5%
META 550.25 ▲1.4%
AMD 521.58 ▼2.1%
AVGO 365.02 ▼3.7%
PLTR 112.93 ▲5.3%
SPCX 153.23 ▲0.2%
TSLA 379.71 ▲1.2%