Thursday, November 12, 2026

AI trade steadies on CoreWeave while founders' bottleneck shifts from building to finding customers

CoreWeave's path-to-profit guidance lifted AI infrastructure stocks after Nvidia shed about $430 billion in two sessions, though regulatory and leadership worries linger. On the go-to-market side, OpenAI opened free business profiles inside ChatGPT, and commentary converges on one point: cheap AI makes products easy to build, so distribution and the customer relationship now hold the value. New York named leaders for its independent AI testing center, and agent-reliability startup Plurai raised a $13M seed.

Sources read
20
Claims that stand
8
Quotes found in the source
17 of 17
Claims struck
0
Corrected or withdrawn
0

We quote it or we drop it. Every claim below carries the source’s own words, checked against the text we read.

Since 11 Nov

What we told you last time, ruled on from today’s evidence.

  • Strengthened
    New York launches first state-run AI testing center1 outlet on 11 Nov, 2 now

    Based at the State University of New York’s Binghamton University, the center will bring together more than 80 scientists from colleges and research institutions across New York state

    Morrisett will lead the center with Jeannette Wing, executive vice president for research at Columbia University.

  • Strengthened
    No-hire business formation jumps in AI-exposed industries2 outlets on 11 Nov, 4 now

    Building an app now takes a weekend, finding customers for it can take years.

    As models become cheaper and more interchangeable, more value may move to the companies that own the customer relationship, workflow and distribution.

  • Gone quiet
    Nothing new today on these. A limit of what we read, not a verdict.

    Australia mandates AI breach reporting after OpenAI agent breach · IBM and Microsoft launch Imagine Enterprise agent platform · Small businesses use AI but have barely touched agents · Graphon raises $35M to fix AI agent memory · OpenAI acquires travel startup Odessia in commerce push · Israeli startup funding reaches $11B, Q3 up 53%

Every correction we have made →

Each mark on a claim’s line is an outlet, placed by when it published: 09:15 to 07:00 UTC.

AI

  1. CoreWeave guidance rescues AI trade after Nvidia sell-off

    CoreWeave shares surged 24.7% after it guided to adjusted profitability in the first half of 2028, lifting the S&P 500 and Nasdaq more than 1% as Nvidia stabilized. Investor confidence in debt-funded AI compute underpins the capacity and pricing every AI startup builds on, and that confidence is swinging day to day.

    earnings from neocloud firm CoreWeave cooled investor concerns about the durability of the AI trade, while Nvidia stabilized after two straight sessions of sharp declines

    3 more receipts

    The company said it would become profitable in the first half of 2028 on an adjusted basis

    Nvidia rises following a rebound in shares of AI companies even as Baird calls for Nvidia CEO Jensen Huang to step down.

    Investors have expressed concerns about more regulation in the AI space after sweeping Democratic victories in the US midterm elections last week.

  2. New York names leaders for independent AI testing center

    Cornell Tech's Greg Morrisett and Columbia's Jeannette Wing will co-lead the New York Center for AI Responsibility and Research at Binghamton, which pulls in more than 80 scientists statewide. A staffed, state-backed evaluator gives regulators an independent check on AI systems, which raises the compliance bar for model and agent builders selling in New York.

    First reported 11 Nov · 2 outlets so far

    Based at the State University of New York’s Binghamton University, the center will bring together more than 80 scientists from colleges and research institutions across New York state

    1 more receipt

    Morrisett will lead the center with Jeannette Wing, executive vice president for research at Columbia University.

  3. HBR: AI task speedups don't equal organizational gains

    Company-reported data from about 7,000 returns at four accounting firms showed AI cut tax-prep time by a third and raised throughput about 50%, but review, client explanation and training costs persisted. Small professional-services firms weighing AI should budget for review and talent-pipeline costs that headline productivity numbers leave out.

    reported that AI tools saved tax practitioners an average of a third of their time on tax preparation and increased throughput by about 50%

    2 more receipts

    AI can dramatically accelerate individual tasks without delivering equivalent gains for the organization, because delegating work creates costs beyond

    The time savings and throughput gains are company-reported.

AI agents that find customers for small businesses and founders

  1. ChatGPT opens free business profiles for customer discovery

    Small businesses can now claim a free profile through the Businesses in ChatGPT feature so their products surface inside the chatbot. ChatGPT becomes a direct, free acquisition channel for small businesses, and founders who ignore AI-search visibility will cede discovery to those who claim it.

    Businesses in ChatGPT is a feature designed to help business owners get their products discovered in the AI chatbot.

  2. Distribution, not building, is now the founder's scarce asset

    Coverage of the solo-founder boom and of AI pricing argues that cheap, interchangeable models move value toward whoever owns the customer relationship and distribution. This defines the market for customer-finding agents: when an app takes a weekend to build, the bottleneck founders will pay to remove is acquisition.

    First reported 10 Nov · 4 outlets so far

    Building an app now takes a weekend, finding customers for it can take years.

    1 more receipt

    As models become cheaper and more interchangeable, more value may move to the companies that own the customer relationship, workflow and distribution.

tech startups

  1. Plurai raises $13M seed to stress-test AI agents

    Plurai raised a $13M Team8-led seed round for a platform that simulates thousands of real-world scenarios before AI agents go live. Agent reliability tooling is drawing capital, a sign buyers won't let customer-facing agents ship without pre-deployment testing.

    Team8-led seed round backs platform that simulates thousands of real-world scenarios before AI agents go live

    1 more receipt

    Plurai raises $13M to catch AI agent failures before customers do

  2. Anthropic reportedly files for a potentially record IPO

    An NPR page headline reports that Anthropic has filed for an IPO that could break records; no further detail appears in today's sources. A public Anthropic would set a valuation benchmark for every AI startup and could reopen the IPO window for the sector.

    AI giant Anthropic files for an IPO that could break records

  3. Legal-tech startups hit buyer 'tool fatigue'

    The FT and Business Insider describe a crowded legal-tech field, with buyers fatigued by tools and the sector large enough to support its own investor summit. Vertical AI startups in saturated categories now compete on distribution and consolidation, not novelty, which raises the bar for new entrants.

    As buyers reach 'tool fatigue', we reflect the breadth of software products available by showcasing eight start-ups tackling four big challenges for law

    1 more receipt

    Legal tech has its own Davos, and it's got much better weather. The LegalTech Fund Summit, which enters its fifth year in Scottsdale, Arizona, next week,

Why it matters

Today's sources point the same way: building with AI is cheap, so customer acquisition is where the advantage sits. OpenAI is turning ChatGPT into a discovery channel for small businesses, and legal tech shows what happens when a category fills with near-identical tools: buyers tire and distribution decides the winners. For founders this means visibility inside AI assistants, plus agents that reliably find and convert customers, matter more than product velocity.

The ground beneath remains unsettled. AI infrastructure sentiment swings on single earnings reports and Nvidia leadership headlines. Post-midterm regulatory risk is now priced in, and New York is staffing an independent evaluator. Plurai's seed round shows capital moving to agent reliability, the precondition for letting agents talk to customers unsupervised.

What to watch

  • Whether Anthropic's IPO filing is confirmed and what valuation it sets for private AI startups.
  • Early adoption data and ranking behaviour for Businesses in ChatGPT profiles, and whether OpenAI adds paid placement.
  • The New York AI center's first evaluation agenda as the RAISE Act's Jan. 1 start approaches.

Sources read

SourceOutletPublished, UTCQuoted
Swatters are using AI voices to deceive 911 dispatchers, and it's only getting worse npr.org01:151×
S&P 500, Nasdaq rise as CoreWeave results ease AI spending worries reuters.com06:153×
The Hidden Costs of Delegating to AI hbr.org05:153×
Greg Morrisett to co-lead new NYS AI safety research center news.cornell.edu06:152×
Introducing usage-based pricing for Amazon SageMaker AI bidirectional streaming aws.amazon.com06:15read, not quoted
AI Walks Into an Architecture Firm… commonedge.org05:15read, not quoted
SpaceX Stock Shrugs Off AI Worries Ahead of Starship Launch barrons.com06:151×
Stony Brook FinTech Workshop Explores AI’s Impact on Finance, Announces New Industry Partnership news.stonybrook.edu06:44read, not quoted
Small Businesses Can Now Claim a Free ChatGPT Profile to Reach Customers. Here's How to Set It Up inc.com22:151×
Solopreneurs in the A.I. Era: Why Solo Founders Are Booming and Who Profits Most trendingtopics.eu07:001×
As AI Gets Cheaper, the Real Business Opportunity Is Shifting. Here's Where the Value Is Moving. entrepreneur.com20:151×
Plurai raises $13M to catch AI agent failures before customers do ynetnews.com15:152×
Databricks to Invest ~$3 Billion Across Europe and the Middle East to Accelerate Enterprise AI Adoption databricks.com09:15read, not quoted
The next big business of AI may be advertising: InMobi founder Naveen Tewari storyboard18.com03:15read, not quoted
How to get started with AI agents, and which are best for specific tasks usa.inquirer.net11:15read, not quoted
Tech Moves: Wizards of the Coasts names president; new execs at Skydance, Pulumi, Marchex and Syndio geekwire.com18:15read, not quoted
Crowded legal-tech market requires navigation ft.com04:151×
Niche work if you can get it businessinsider.com20:151×
North Dakota rode a boom to the top. Now it has to build what’s next. technical.ly13:15read, not quoted
Mapping Israel’s WasteTech sector: 80 startups up against a 76% landfill problem calcalistech.com19:15read, not quoted

AI Builder’s Edge

Apiiro found Claude Code's .claude/settings.local.json committed to more than 104,000 public GitHub repos. That file stores approved shell commands verbatim, and 1,122 copies held live credentials. OpenAI is acquiring Pierre Computer to fold its Code Storage into Codex, joining Cursor and ex-GitHub CEO Thomas Dohmke in building AI-first alternatives to GitHub. Today's sources carry no substantive signal on AI video or creator tooling.

Sources read
6
Claims that stand
3
Quotes found in the source
8 of 8
Claims struck
0

We quote it or we drop it. Every claim below carries the source’s own words, checked against the text we read.

Each mark on a claim’s line is an outlet, placed by when it published: 07:16 to 16:16 UTC.

  1. Claude Code's 'don't ask again' writes secrets into committed config

    Apiiro found over 104,000 public copies of .claude/settings.local.json, 1,122 containing live secrets, because approving a command with 'Yes, and don't ask again' saves the full command text, including inline credentials, in plain text. Builders who pass tokens inline in shell commands and approve them permanently should gitignore .claude/settings.local.json, scrub it from Git history and rotate any key that touched it.

    Apiiro found more than 104,000 copies of this settings file sitting in public GitHub repositories. Of those, 1,122 contained real, active secrets.

    2 more receipts

    When a developer approves a shell command using that option, Claude Code writes the full text of that command into the .claude/settings.local.json file

    By examining Git history and other related files, the researchers identified 2,539 distinct credentials spread across nearly 2,000 repositories.

  2. OpenAI acquires Pierre Computer to bring Code Storage into Codex

    OpenAI told Axios it will acquire Pierre Computer, a developer tools startup already embedded in Codex, and integrate Pierre's Code Storage function into Codex; financial terms were not disclosed. Codex is moving from coding agent toward owning where code is stored, which could change where Codex users keep and version agent-written code.

    OpenAI will acquire Pierre Computer, a developer tools startup whose technology is already embedded in Codex, the company told Axios exclusively.

    1 more receipt

    OpenAI plans to integrate Pierre's Code Storage function into Codex, where Pierre's technology is already used.

  3. AI-first GitHub replacements multiply as agents write more code

    Cursor, now owned by SpaceX, launched the GitHub rival Origin in August, and former GitHub CEO Thomas Dohmke announced a competing product last month, as critics fault GitHub's reliability and human-centric design. Repo hosting, the default layer under every coding workflow, is now contested, and builders may soon choose a repo platform based on which agent stack they use.

    Cursor, now owned by Elon Musk's SpaceX, launched GitHub rival Origin in August.

    2 more receipts

    Former GitHub CEO Thomas Dohmke, who launched a new coding startup, announced a competing product last month.

    it was built for humans, but so much software is now written by AI that a new, AI-first GitHub replacement would be better.

Why it matters

Both stories point the same way: agents now produce artifacts that land straight in shared infrastructure. The Claude Code leak shows that convenience features such as permanent command approval create new files that ordinary .gitignore habits miss, and Git history keeps the damage retrievable after a fix. Any agent that persists approved actions to disk carries the same class of risk.

OpenAI buying storage technology for Codex, Cursor's Origin and Dohmke's product show labs and toolmakers competing for the repository layer, not just the editor or the model. Microsoft owns nearly a quarter of OpenAI, so this is competition with a partner. For builders, the near-term effect is fragmentation: the coding agent you pick may increasingly determine where your code lives.

What to watch

  • Whether Anthropic changes Claude Code to redact credentials, gitignore settings.local.json by default, or warn before saving inline secrets.
  • When Pierre's Code Storage appears inside Codex, and whether Codex begins to work as a standalone alternative to GitHub.
  • Whether GitHub responds to the AI-first rivals with agent-native features or reliability fixes.

Know someone watching this? Pass it on.

20 sources, at least 20 minutes of reading. This briefing took about 3. You’re done.

Counted from the 7 articles we could read in full, at 230 words a minute; 13 more sat behind a paywall or gave us only a headline.