ENFR
8news

Tech • IA • Crypto

TodayBriefingVideosTop 24hArchivesFavoritesTopics

SpaceX $2.28T IPO, Claude Fable 5 Ban, Google Diffusion Gemma

AISunday, June 14, 2026· 11 videos

Briefing

Audio player
0:00 / 0:00

SpaceX IPO hits $2.28T valuation

SpaceX debuted publicly at $135 per share, opening near $150 and climbing to roughly $167–$172 on day one. The move delivered a 20–25% gain, pushing valuation to about $2.28 trillion. Demand vastly exceeded allocation, with many investors receiving minimal shares. Analysts view the controlled surge as a sign of strong pricing discipline and long-term institutional confidence.

U.S. bans Anthropic Fable 5

Anthropic’s Claude Fable 5 and Mythos 5 were abruptly shut down following a June 12, 2026 U.S. government order. The restriction applied globally and even impacted foreign nationals داخل the U.S., forcing a full service shutdown. Officials cited national security concerns, reportedly tied to jailbreak vulnerabilities. Anthropic disputes the severity, arguing similar risks exist across competing models.

Google unveils Diffusion Gemma model

Google introduced Diffusion Gemma, a 26B MoE text model that generates outputs via iterative refinement instead of token-by-token prediction. It operates on a 256-token canvas, enabling full-text revisions mid-generation. The system delivers over 1,000 tokens/sec on H100 GPUs and runs locally with ~18GB VRAM. Google positions it for structured tasks like coding, OCR, and editing rather than pure writing quality.

Microsoft launches MAI Thinking One

Microsoft revealed its in-house MI model family, led by MAI Thinking One (35B parameters). The models emphasize coding and enterprise performance over frontier-scale capability. This marks a strategic shift away from reliance on OpenAI, toward vertical integration across cloud and AI infrastructure. The move aligns with a broader trend favoring cheaper, “good enough” models at scale.

Musk folds xAI into SpaceX

Elon Musk dissolved xAI, merging it into SpaceX AI ahead of the IPO. The Grok model becomes part of a broader infrastructure strategy rather than a standalone competitor. xAI lagged far behind rivals, with roughly $500M revenue vs Anthropic’s $47B run rate and weak enterprise adoption. The pivot signals a shift from model competition to compute and platform dominance.

Claude skills redefine AI agents

Anthropic’s Claude skills introduce a modular alternative to complex agent systems. Each skill is a structured set of Markdown files defining tasks, context, and outputs without code. Tools like SkillSmith can generate these in minutes, lowering development barriers. The approach reflects a broader shift toward simpler, composable AI workflows.

French gov breach exposes 70,000 users

France’s public-sector messaging platform suffered a breach affecting about 70,000 agents. Attackers claim access to 13.5GB of data, including 640,000 messages across 800 channels. Officials say under 9% of data was compromised, but sensitive interministerial content was included. The incident appears driven by social engineering, highlighting human—not technical—failures.

Odysseus unifies ChatGPT and Claude

Odysseus, backed by PewDiePie, aggregates models like ChatGPT, Gemini, Claude, and Mistral into one interface. It supports both local deployment via Docker and cloud APIs, enabling flexible workflows. Users can run side-by-side model comparisons and share persistent memory across systems. The platform reflects growing demand for multi-model orchestration and control.

Videos covered

Previous briefings · AI