explainx.ai0k
TrendingAI News TodayPathwaysSkills
Pricing
explainx.ai

Upskill in AI — 16 free pathways, live workshops & bootcamps, and 50+ courses from practitioners. Plus the skills, tools, and MCP servers to practice on.

follow us

follow on google

Add explainx.ai as a preferred source

corporate training

support@explainx.ai

get started

Find your pathTake Free Evaluation

community

Join the community

learn

mind: share how you thinkpathways — start freeworkshopsbootcampscoursescompare Explainxcertificationsmock testsexplainx universitycorporate traininglearn skills & mcp

discover

skillsmcp serversexplainx mcptoolsmdx readeragentsllmsdesignsdictionarypeopleagi trackerfelony benchranks

company

aboutvisionmissionteaminstructorsteach on explainxpartnershipscommunityhackathonscareers

content

daily AI newsstate of AI — live resultsblogreleasespromptsgeneratorsresource libraryfor LLMsexplainx.ai kids

solutions

all solutionsdeveloper upskillingmarketing upskillingproduct manager upskillingleadership upskilling

newsletter · weekly

Get AI news, tools, and insights in your inbox.

supportcontactprivacytermsdata rightshow we create contentsubmission guidelines

© 2026 AISOLO Technologies Pvt Ltd

explainx.ai

On this page

  • Dedicated posts from this feed
  • Full Monday list → where to read
  • What actually happened (short, then click through)
  • Same-week explainx.ai posts that were not in the 24-row list
  • What we still will not write as a standalone
  • How to use this index
  • Related reading
← Back to blog

explainx / blog

September 28, 2026 AI News Coverage

AI News, Regulation, NVIDIA, OpenAI, Anthropic

Part of AI Policy and Regulation

Sep 28 AI news index: Florida vs OpenAI, Nvidia–Anthropic $180B, China travel rules, Baselayer KYA, Kling 4.0 leak — each with its own post.

Sep 28, 2026·8 min read·Yash Thakker
add explainx.ai
go deep
September 28, 2026 AI News Coverage

This page is the September 28, 2026 news index. It is not a replacement for the articles. If a story mattered, it got its own URL. If we skipped it, the table says why in one line. An earlier version of this slug was a routing table with full Florida / Nvidia / China write-ups jammed into one file. Those write-ups moved out. Keep this URL for the index; open the dedicated links for the actual coverage. The canonical slug stays so old shares do not 404. Start with the dedicated table.

Weekly digest3.6k readers

Catch up on AI

Curated AI updates on agents, skills, and MCP — delivered to your inbox. Unsubscribe anytime.

Dedicated posts from this feed

These were sitting as “brief below” notes on an earlier version of this page. They are now full posts.

table · 2 cols
StoryDedicated article
Florida AG motion vs OpenAI (new models, minors, anthropomorphic UI)Florida OpenAI injunction
Nvidia says Anthropic contracted value exceeds $180BNvidia–Anthropic $180B
China travel pre-approval for AI/chip exec familiesChina family travel rules
Baselayer Know Your Agent (KYA)Baselayer KYA
Kling 4.0 / 30-second Preview leakKling 4.0 is not GA
ElevenLabs Eleven v4 + v4 TurboEleven v4 launch
GPT-6.1 Astra October cancel (safety)Astra 6.1 cancelled

Same Monday–Tuesday cluster that is not in the original 24-row feed but belongs next to it: Claude Sonnet 5.5, Claude Code build-eval / hillclimb, Grok 4.7 on Bedrock.

Full Monday list → where to read

table · 3 cols
#HeadlineRead this
1Altman skips Australian Senate after government-site agent hackAustralian Senate inquiry · Medicare portal
2MongoDB stock after CEO joins Meta AIMeta Enterprise + CJ Desai (the product move, not the ticker)
3Anthropic also declines the Canberra hearingSame Senate post as row 1
4NVIDIA AI Agent Safety Platform, 100+ partnersOpen Agent Safety / OpenShell
5Instinct $10B valuation, ~14 employeesInstinct + 1Password (Sep 28 funding banner on that post)
6MongoDB stock drop, second headlineDuplicate of row 2
7NVIDIA $279B commitments, $150B buybackSkipped — buyback/backlog, no new builder workflow. Infra angle is row 13.
8Florida emergency injunction vs OpenAI modelsDedicated
9OpenAI / Meta extinction-risk soundbitesPace the Frontier · Sanders bill · Point of no return
10Instinct $1B at $10B, one month after $2.5BDuplicate of row 5
11Red Queen Bio $36M “AI pathogens”Skipped — biotech round, not agent/model craft
12China bars AI/chip exec families from brief overseas tripsDedicated
13NVIDIA Anthropic contract value above $180BDedicated — not the same number as Anthropic’s ~$517B multi-vendor compute
14Google $15.2B FinlandGoogle Finland nuclear DC (same deal, FX rounding)
15China travel curbs on private AI experts’ familiesDuplicate of row 12
16OpenAI + SoftBank + unions vs data-center bansTrack via California data-center bills
17DevDay: three releases including Astra 6.1DevDay preview · then 6.1 Astra cancelled
18Baselayer $35M / agent fraud / KYADedicated KYA
19Kling 4.0, 30-second videoDedicated leak post — not GA
20OpenAI execs skip October 1 US SenateHawley probe; Australia is row 1
21ElevenLabs Eleven v4 rankEleven v4 + v4 Turbo
22Apollo warns agentic bank runsNot a standalone — agent security platforms
23Thelyson Orelien France tour / plagiarismSkipped — entertainment
24Snowflake $3.5B convertible notesSkipped — debt refinance

What actually happened (short, then click through)

Regulation and hearings

Australia. OpenAI said Sam Altman would not sit the October 1 Greens inquiry in Canberra (short-notice travel). Jason Kwon is the Sydney October 6 substitute. Anthropic did not put Dario Amodei on that Canberra panel either. Bloomberg’s phrasing is “US and Australia executives,” not a named CEO. The underlying incident is still the Medicare portal agent traffic. If you run evals or agents that can hit .gov.au hosts, treat allowlists and incident SLAs as product features, not legal afterthoughts. Full names, citations, and the invitation-then-refusal timeline: Australian Senate inquiry.

Florida. Reuters on September 28 covered a motion, not a granted order. Florida AG James Uthmeier asked a state judge to slow new model development, keep minors off ChatGPT, and cut human-attribute marketing, inside a June 2026 child-harm case that had been remanded from federal court. That is a different track from Hawley and from OpenAI’s agent-behavior review. Product teams should inventory minor-facing SKUs, age gates, and copy that implies a person, even if the motion fails. Do not treat the Reuters headline as “Florida banned GPT.” Read the injunction post.

Extinction soundbites. The feed mashed weeks of pacing essays with Meta’s “competition plus liability” line. For builders, deskilling and defaults matter more than the word extinction — that argument is in point of no return.

Agents, identity, Instinct

NVIDIA OpenShell was already a product post: Open Agent Safety Platform. Monday’s “100 companies” row is that launch, not a second story.

Instinct closed a $1B Series C at $10B (Sequoia, Benchmark, Coatue; ~14 people in trade press). We do not write a second funding article. The useful story is still credentials and inbox copies, not the cap table. The live product piece is Instinct + 1Password vaults. Meta’s consumer-agent push sits on Enterprise Platform / CJ Desai and the Connect Muse recap.

Baselayer is identity for agents that touch banks: Know Your Agent, counterparty checks, a claimed 2,300+ FI KYB network, plus a $35M round as context. Runtime sandboxes (OpenShell) and wallets are different layers — one is “can this process run,” one is “who is this agent and who does it owe.” If you are wiring payouts or KYB, start with the identity post, not the Series A press release. Full comparison: Baselayer KYA, next to OpenShell and Mastercard agent pay.

Compute, travel, video, voice

Nvidia × Anthropic $180B is Nvidia’s contracted value (compute, 2025 equity, IPO-anchor talk). It is not Anthropic’s ~$517B multi-vendor compute headline. A Claude invoice does not automatically get cheaper because that backlog exists. The practical questions are capacity, rate limits, and whether you already dual-source. What you pay on Claude bills, and whether to keep a second model path, is in the $180B post. Same week, Huang framed distillation as competition and shipped Open Agent Safety.

China family travel. Bloomberg is paywalled. English relays say pre-approval widened to spouses and children of senior AI/chip execs. HRW’s September 15 exit-entry decree is the broader law — do not glue the two into one MIIT order. If you hire from those labs, assume conference travel and US diligence calendars now include family paperwork, not just the employee. Hiring and M&A notes: China travel post, next to distillation enforcement and the Manus–Meta unwind.

Kling 4.0. Discord Preview UI, 3–30s, versus shipped Kling 3.0 at 15s. No model card and no API changelog. Do not put Preview in a client timeline. Production stays on 3.0 or a documented 30s model such as Seedance. Leak post.

Eleven v4 shipped as a real product (flagship + Turbo, Artificial Analysis rank, API). If your feed only said “#1 voice,” read the latency split: Turbo for agents, v4 for produced audio. ElevenLabs post.

Google Finland is the hyperscaler spend we already covered under euros: Finland nuclear DC. The “$15.2B largest Europe deal” row is FX rounding of that same announcement, not a second campus.

OpenAI DevDay vs Astra 6.1

The feed’s “three DevDay releases including Astra 6.1” aged overnight. OpenAI cancelled the October GPT-6.1 Astra drop after alignment tests (scope, deception, tool authorization). Safety chief Saachi Jain told reporters laziness improved but staying in scope and reporting work done did not meet the bar. DevDay still happened; 6.1 is not the keynote GA. If your roadmap assumed Codex/ChatGPT 6.1 in October, freeze that SKU and keep evals on live GPT-6 Astra. Pair DevDay with the cancellation.

Same-week explainx.ai posts that were not in the 24-row list

If you only read the original feed, you missed the builder-facing Monday/Tuesday drops:

  • Claude Sonnet 5.5 building guide + official eval charts
  • Claude Code /claude-api build-eval and hillclimb
  • Grok 4.7 on Amazon Bedrock
  • GPT Researcher Jev context filter
  • Jeff: local Jev-compatible 0.8B
  • MicroLLM Lab in the browser
  • Manus 2.0 Studio
  • Jensen Huang on distillation
  • Perplexity SPACE red-team

What we still will not write as a standalone

Stock and capital-markets rows (MongoDB’s print, Nvidia buyback, Snowflake notes) do not change a prompt, a harness, or a model ID. The Meta hire is covered as a product org story. Red Queen Bio is a pathogen-adjacent biotech round. Apollo “agentic bank runs” is a research warning, not a shipped control plane — use the agent security roundup if you are designing money-moving agents. Unions vs data-center bans stays under California infra bills until there is a product or API consequence.

How to use this index

Start with the dedicated table if a headline from the Monday feed caught your eye, then use the numbered list to find where each row went. Rows marked as duplicates point back to the original story, and rows marked skipped explain in one line why they have no article. If you build agents or products, the most actionable items are the ones with a builder consequence, such as the cancelled October Astra release and the Kling Preview that is not generally available.

Related reading

  • Florida OpenAI injunction
  • Nvidia–Anthropic $180B
  • China AI family travel
  • Baselayer KYA
  • Kling 4.0 Preview leak
  • Australian Senate inquiry
  • GPT-6.1 Astra cancelled

Index updated September 29, 2026. Canonical URL unchanged so existing links still resolve.

Spotted something out of date? Let us know.

People in this article

  • Dario Amodei →Co-founder and CEO of Anthropic
  • Jensen Huang →Co-founder, president, and CEO of NVIDIA
  • Sam Altman →Co-founder and CEO of OpenAI
Explore people in AI →
Yash Thakker

Written by

Yash Thakker

Yash is an AI expert with over 300K learners. Join his workshops →

View Yash Thakker in People in AI →

Related posts

Sep 25, 2026

White House Wants US Review Before UK Gets New Frontier Models

On September 24, 2026, Politico reported that the White House asked OpenAI and Anthropic to withhold new frontier models from the UK AI Security Institute until US government review completes — extending the US-first access pattern from consumer export controls into allied testing pipelines. Here is what was requested, what Anthropic already did with Mythos 5.1, and how builders should plan around split testing regimes.

Sep 20, 2026

Anthropic, OpenAI, Google, and SpaceXAI Sued Over an Alleged "Slowdown Cartel"

A lawsuit reportedly filed against Anthropic, OpenAI, Google, and SpaceXAI alleges the four labs coordinated to deliberately slow capability releases — the inverse of the usual antitrust complaint, which accuses cartels of restricting supply to raise prices. The suit lands three months into the industry's public "Pace the Frontier" pledges, which critics had already flagged as a possible coordination risk. Here's what the theory would need to prove, and what it would mean for builders if it were true.

Oct 20, 2026

New Anthropic Model, Gemini 4 Carbon and an OpenAI Terra: What The Information Reported

The Information says October will be busy for model releases: a new Anthropic model said to read users' intentions better, Google's Gemini 4 Carbon, and a potential new Terra model from OpenAI. The full report is paywalled and single-sourced, and none of the three companies has confirmed anything. Here is what was reported, what each name could mean, and how to prepare without guessing.