Latest / DX Today | No-Hype Podcast & News About AI & DX / AI News Today: GPT 6 Goes Free for Every ChatGPT User

Transcript

AI News Today: GPT 6 Goes Free for Every ChatGPT User · DX Today | No-Hype Podcast & News About AI & DX

• 0:00
  SPEAKER_05: It's Thursday, October 8, 2026.
• 0:03
  SPEAKER_05: You're listening to the DX Today AI Daily Brief.
• 0:07
  SPEAKER_05: Today, OpenAI puts GPT-6 in front of every ChatGPT user.
• 0:12
  SPEAKER_05: Samsung posts a record$80 billion a quarter on AI memory demand, and Broadcom seeks more than$50 billion to finance OpenAI's custom chip.
• 0:23
  SPEAKER_05: Let's get into it.
• 0:25
  SPEAKER_00: OpenAI has begun rolling out GPT-6 inside ChatGPT, and this time free users are included.
• 0:32
  SPEAKER_00: The company says the model reached plus pro business and enterprise subscribers on Wednesday, with free and go users following starting today.
• 0:40
  SPEAKER_00: Alongside it comes a feature OpenAI calls Intelligent UI.
• 0:45
  SPEAKER_00: Instead of answering only in text, ChatGPT can now decide when a chart, a diagram, a form, tappable buttons, or a small working tool would serve the question better and build that interface right inside the conversation.
• 1:00
  SPEAKER_00: GPT-6 can also begin responding while it is still reasoning.
• 1:04
  SPEAKER_00: OpenAI says paid and free plans run different versions of the model.
• 1:08
  SPEAKER_00: The move puts the company's newest model in front of its entire user base within two days and raises the bar for what a chat assistant's answer can look like.
• 1:17
  SPEAKER_02: Next, a price war at the small end.
• 1:20
  SPEAKER_02: Anthropic has released Claude Haiku 5.5, its fastest and least expensive model, and its third, Claude 5.5 release in roughly a month.
• 1:32
  SPEAKER_02: The headline is price.
• 1:34
  SPEAKER_02: For prompts under 100,000 tokens, which Anthropic says made up about 90% of requests to its previous Haiku model, input costs 10 cents per million tokens and output costs 50 cents.
• 1:49
  SPEAKER_02: That is roughly a 90% cut.
• 1:51
  SPEAKER_02: Longer prompts are priced higher at a smaller discount.
• 1:55
  SPEAKER_02: Anthropic says the new model costs about 75% less to run on average than Haiku 4.5.
• 2:03
  SPEAKER_02: The launch targets high volume work where cost per request matters most, and it lands as the company prepares for a planned public offering, with every release now read as part of that story.
• 2:16
  SPEAKER_04: Now to sew, Samsung Electronics says its third quarter operating profit hit a record, about$107.4 trillion1, or roughly$80 billion.
• 2:27
  SPEAKER_04: That preliminary estimate is up more than 780% from a year earlier, and it tops the record set just last quarter.
• 2:35
  SPEAKER_04: Revenue is expected near$195 trillion, one, more than double a year ago.
• 2:41
  SPEAKER_04: It is the first time a South Korean company has posted a quarterly operating profit above$100 trillion1.
• 2:48
  SPEAKER_04: The driver is the AI memory boom, as demand for the high bandwidth memory used in AI accelerators lifts chip prices across the industry.
• 2:57
  SPEAKER_04: Samsung's guidance did not break out results by division.
• 3:00
  SPEAKER_04: For investors, it is fresh evidence that the AI infrastructure buildout is still accelerating.
• 3:06
  SPEAKER_04: Following the money.
• 3:08
  SPEAKER_01: Broadcom is working to arrange more than$50 billion in financing for the custom AI chip it is building with OpenAI, according to the Wall Street Journal.
• 3:17
  SPEAKER_01: The paper reports that lenders, including Apollo Global Management and Blackstone, are in talks, and that the discussions are early, so the size could still change.
• 3:27
  SPEAKER_01: Broadcom is not alone.
• 3:29
  SPEAKER_01: Oracle is reportedly talking with Apollo and Goldman Sachs about funding its own chip purchases, and SpaceX is seeking a large loan to buy NVIDIA processors.
• 3:40
  SPEAKER_01: Together, the deals show how the cost of AI infrastructure is moving off corporate balance sheets and into private credit markets.
• 3:48
  SPEAKER_01: For OpenAI, the financing would help turn its custom silicon plan into hardware at scale.
• 3:54
  SPEAKER_01: The report cites people familiar with the talks, and terms have not been finalized.
• 3:59
  SPEAKER_05: To Windows and the desktop.
• 4:01
  SPEAKER_03: Microsoft used a Windows and Surface event in San Francisco on Wednesday to lay out what it calls hybrid intelligence.
• 4:09
  SPEAKER_03: AI agents that run on the PC when they can and reach to the cloud when they need more power.
• 4:14
  SPEAKER_03: Chief Executive Satya Nadella shared the stage with NVIDIA's Jensen Huang.
• 4:20
  SPEAKER_03: Microsoft made its execution containers generally available on Windows 11, a security layer that lets administrators control which files and networks an agent can touch, and that logs agent actions separately from the user's own.
• 4:33
  SPEAKER_03: The company also showed a compressed version of its coding model, MAI, code 1.1 Flash, running locally on PCs, and it opened pre-orders for the Surface Laptop Ultra with Nvidia's RTX Spark chip, starting at$2,599.
• 4:52
  SPEAKER_03: The pitch, cheaper, more private AI work on the desk.
• 4:56
  SPEAKER_05: Now, an agent startup reborn.
• 4:59
  SPEAKER_00: Manus, the Chinese-built AI agent startup, has raised more than$500 million in its first funding round since its acquisition by Meta was unwound.
• 5:09
  SPEAKER_00: Parent company Butterfly Effects said today the round was led jointly by Boyu Capital and IDG Capital, with existing investors including Tencent and Gen Fund also taking part.
• 5:20
  SPEAKER_00: The company did not disclose a valuation.
• 5:23
  SPEAKER_00: Meta had agreed to buy MANIS for more than$2 billion, but the deal was reversed earlier this year on orders from Beijing, and MANIS resumed independent operations.
• 5:33
  SPEAKER_00: Its general purpose agents carry out multi-step tasks on their own, from research to building software.
• 5:39
  SPEAKER_00: The raise signals that investors still see value in MANIS as a standalone company, even after an international deal collapsed under regulatory pressure.
• 5:47
  SPEAKER_05: A new check for synthetic media.
• 5:49
  SPEAKER_02: Google has opened its synth ID detector to everyone.
• 5:54
  SPEAKER_02: The free website lets anyone upload an image, a video, or an audio file, and check whether it carries synth ID, the invisible watermark that Google's AI tools embed in what they create.
• 6:08
  SPEAKER_02: The bigger change is reach.
• 6:11
  SPEAKER_02: The detector now also recognizes watermarked content from Google's partners, including OpenAI, Nvidia, and Cacao, and Google says Apple support is coming soon.
• 6:24
  SPEAKER_02: The tool is available globally in English for now.
• 6:28
  SPEAKER_02: There is an important limit.
• 6:29
  SPEAKER_02: It detects the watermark, not AI content in general.
• 6:33
  SPEAKER_02: So media from tools that do not use synth ID will not be flagged.
• 6:38
  SPEAKER_02: Still, as synthetic images and voices spread across social feeds, a shared industry-wide check is a meaningful step toward telling what is real.
• 6:49
  SPEAKER_04: Turning to the workforce, Amazon has made another round of job cuts, this time mostly in its stores division, which runs the main shopping website.
• 6:57
  SPEAKER_04: The company confirmed on Wednesday that a small number of roles were affected.
• 7:01
  SPEAKER_04: A person familiar with the matter told Reuters that fewer than 1,000 white-collar employees lost their jobs across the United States, India, and the United Kingdom, with reports pointing to customer service, marketplace support, and engineering roles.
• 7:17
  SPEAKER_04: The timing drew attention, landing in the middle of prime big deal days, one of Amazon's biggest sales events.
• 7:24
  SPEAKER_04: The cuts follow a much larger reduction of about 30,000 corporate jobs, and they come as Amazon pours record sums into AI infrastructure.
• 7:32
  SPEAKER_04: It fits a broader pattern across big tech, with spending shifting toward data centers, while headcount in older functions shrinks.
• 7:40
  SPEAKER_04: Back to data centers.
• 7:41
  SPEAKER_01: Data center developer applied digital reported fiscal first quarter revenue of$341.9 million, up 322% from a year earlier and far above analyst expectations.
• 7:55
  SPEAKER_01: Most of the jump came from fitting out and renting facilities for AI and high-performance computing customers.
• 8:01
  SPEAKER_01: The company said its leases now represent about$36 billion in contracted revenue over their initial terms, potentially rising to about$86 billion with renewals.
• 8:12
  SPEAKER_01: It also signed a 210-megawatt 15-year lease at its Delta Forge 2 campus, worth roughly$5.2 billion.
• 8:21
  SPEAKER_01: Applied Digital still posted a net loss driven by interest and non-cash charges, but its adjusted loss narrowed sharply.
• 8:29
  SPEAKER_01: Shares jumped in early trading after the report.
• 8:31
  SPEAKER_05: Next, open source agents.
• 8:33
  SPEAKER_03: News research has confirmed a$90 million Series B at a valuation of$1.5 billion.
• 8:41
  SPEAKER_03: The round was led by Robot Ventures, with Nvidia, Union Square Ventures, Menlo Ventures, and Samsung among the participants.
• 8:50
  SPEAKER_03: News builds Hermes Agent, an open source AI assistant that can search the web, edit files, and write code, and that has become popular with developers.
• 9:01
  SPEAKER_03: Alongside the funding, the three-year-old company launched Hermes for Businesses, which lets companies run custom agents for multi-step workflows without their data leaving their systems.
• 9:12
  SPEAKER_03: The model is open core.
• 9:14
  SPEAKER_03: The agent software stays free, while businesses pay for hosting, security, management, and support.
• 9:21
  SPEAKER_03: The RAISE brings total funding to$158 million and tests whether enterprises will pay for an open agent rather than a closed one.
• 9:30
  SPEAKER_05: Now robots need data too.
• 9:32
  SPEAKER_00: Mecca, a two-year-old Toronto robotics data startup, has raised a$60 million Series B led by Sequoia Capital with Nvidia, Microsoft's Venture Fund M12, Qualcomm, and Samsung also investing.
• 9:46
  SPEAKER_00: Mecca pays people to wear body sensors while they perform everyday tasks, then turns those recordings into training data for humanoid and other robots, including motion tracking and three-dimensional reconstruction.
• 9:59
  SPEAKER_00: The company says it will use the money to scale its data infrastructure, expand research, and begin commercial robot deployments.
• 10:06
  SPEAKER_00: The bet reflects a basic problem in physical AI.
• 10:09
  SPEAKER_00: Unlike text, human motion data cannot be scraped from the internet, so it has to be collected.
• 10:14
  SPEAKER_00: Whoever builds the largest and cleanest supply of it may hold a real advantage as robot makers race to train general purpose machines.
• 10:21
  SPEAKER_02: And finally, to New Delhi.
• 10:23
  SPEAKER_02: India's information technology minister, Ashwini Vaishnar, said today that the government will release a consultation paper on artificial intelligence regulation within a month.
• 10:34
  SPEAKER_02: He said the time has come for India to have such a framework, and that the paper will form the basis for the country's approach to governing the technology.
• 10:44
  SPEAKER_02: Key areas include AI safety, deep fakes, and what he called a human-first approach, along with skilling and making sure the benefits are widely shared.
• 10:57
  SPEAKER_02: Humanity, he said, has to be the most important part of this.
• 11:01
  SPEAKER_02: The paper will seek views from the public and industry on how AI should be governed before any rules are drafted.
• 11:09
  SPEAKER_02: It marks a concrete step toward formal AI regulation in one of the world's largest and fastest growing digital markets.
• 11:18
  SPEAKER_05: That's your briefing for Thursday, October 8, 2026.
• 11:22
  SPEAKER_05: For DX Today, stay curious.