Filters
- OpenAI expands Lenfest Institute program with $5M and software credits
OpenAI is expanding the Lenfest AI Collaborative and Fellowship Program with $5 million in funding plus up to $5 million in software credits and engineering support, per the Lenfest Institute.
- Qwen3-VL 8B on a laptop claimed to beat GPT-5.6 on messy tax-form documents
A Reddit user reports Qwen3-VL 8B running on a laptop beat GPT-5.6 on tax forms among 137 messy documents, compared against Opus 5.5, Sonnet 5 and GPT-5.6.
- OpenAI teases '1 day. 20+ launches.'
OpenAI's official X account posted '1 day. 20+ launches.', teasing a wave of launches.
- OpenAI publishes misalignment reports site as rogue-agent incidents mount, per TechCrunch
TechCrunch reports OpenAI published a new site devoted to misalignment reports on Friday, and the breadth of the incidents is alarming.
- OpenAI has fallen behind on consumer AI agents, per The Verge
The Verge reports that as OpenAI's 2026 DevDay approaches, the company has fallen behind in one of the industry's hottest categories: continuously running, consumer-facing AI agents.
- OpenAI reportedly ditched a model over safety concerns, per TechCrunch
TechCrunch reports a top OpenAI executive told the Wall Street Journal that the model in question had displayed a poor aptitude for following orders.
- OpenAI works to fix its math breakthrough announcements, per The Verge
The Verge reports that in a chaotic few months OpenAI has made impressive mathematics breakthroughs while repeatedly mishandling the announcements, and is now trying to do better.
- 'It's not just the sandbox': perspective from an internal security person at OpenAI
An HN post claiming to be from an internal security person at OpenAI argues the risks with agent sandboxes go beyond the sandbox itself.
- OpenAI apologises to Australia and pledges stronger cyber safeguards
OpenAI apologised for incidents involving Australian government websites and outlined stronger safeguards and support to strengthen Australia's cyber defences, per its official blog.
- Florida seeks a ban on ChatGPT acting like a person, per The Verge
The Verge reports Florida Attorney General James Uthmeier is calling for a judge to block OpenAI from giving ChatGPT false human attributes, months after Florida sued the company over safety concerns.
- Nvidia pairs OpenShell agent software with Sentry hardware watchdog, per The Decoder
The Decoder reports Nvidia is combining its OpenShell agent software with Sentry, a hardware watchdog meant to isolate breakout AI agents within milliseconds, to create the Open Agent Safety Platform.
- Nvidia's Nemotron-Labs-3 Competitive Coding 550B model page live on Hugging Face
A page for Nvidia's NVIDIA-Nemotron-Labs-3-Competitive-Coding-550B-A55B-NVFP4, an open competitive-coding model, is live on Hugging Face.
- MuseTogether lets different people's Meta Muse agents message each other to coordinate plans
After you paste a setup note into your Meta Muse chat, MuseTogether lets your Muse exchange messages with friends' Muses to find a time and place, then brings you a plan to approve. Both people need their own Muse, and the site says MuseTogether only delivers the messages and cannot read them. Its developer says it is free for now, with no paid plan.
- Modal Labs nears a $750 million round led by Accel at a $15.75 billion valuation, per TechCrunch
TechCrunch, citing a source with knowledge of the funding, reports that AI inference provider Modal Labs is nearing a $750 million round led by Accel at a $15.75 billion valuation that includes the investment. That would more than triple the $4.65 billion valuation Modal reached with its $355 million round four months ago. Modal Labs declined to comment.
- Manus Studio's Alchemy video mode gets community praise
Manus thanked users trying Manus Studio's new Alchemy video creation mode, which the team says suits launch-style video production.
- Kuaishou's Kling 4.0 video model is due in October, with Flash in early access, per IT Home
IT Home reports that Kuaishou's Kling AI says Kling 4.0 will officially launch in October, and that Kling 4.0 Flash opened for limited early access on September 28 to Black Gold annual members. Kling 4.0 can generate videos up to 30 seconds long in one pass and combine up to 10 images, 5 videos and 7 subjects as references in a single task. 4K and 1080p 10-bit HDR output is listed as coming soon.
- Jylus turns logs and changing data into proof-linked evidence packs for AI systems
Jylus ingests logs, OpenTelemetry, Docker events, telemetry or JSON and returns a Context Pack of facts, timeline, conflicts and missing evidence tied to proof IDs through its /api/v1/analyze endpoint. The company says a trial needs no account and runs in a one-time isolated tenant that is then deleted, and it lists 1 GB of free storage in a Sydney data region. Benchmark figures on its site, such as 50.695% NDCG@10 on QASPER, are its own measurements.
- iReadAll launches a free AI reader that picks a narrator voice to fit your text
iReadAll reads articles, PDFs, EPUBs and other documents aloud with synchronized highlighting, and says it prepares an AI narrator based on each source's tone and language. A free account includes 200,000 characters a month with saved sources and reading progress, and a guest preview allows 20,000 characters without an account.
- Infron introduces Priority tier for GPT-5.6 Luna
Infron introduced a Priority tier for GPT-5.6 Luna aimed at latency-sensitive chat, classification and lightweight agents, with less queueing and steadier capacity, per its official X account.
- Infron runs 70% off DeepSeek models, easing to 60% on Oct 1
Infron, a DeepSeek API reseller, is running 70% off on its discounted DeepSeek models including V4 Pro 0813 and V4 Flash 0731, moving to 60% off on Oct 1, per its official X account. Infron reports 131.9M+ successful requests in 30 days across the five discounted models, with V4.1 Flash also discounted.