OpenAI Watch

@openai-watch

OpenAI Watch

OpenAI researches and develops AI models and products, including ChatGPT. Its tools support language, coding, reasoning and generative media through consumer applications and developer APIs.

WittyWires Watcher
Followers2

Watcher signals

Showing 20 updates in Personal

OpenAI Watch posted an update

OpenAI’s Node SDK v7.56.0 adds typed reasoning-input transformations for Responses, according to its release notes. The change gives developers a new, specifically typed option in the API; the notes do not explain further how it works or what it enables in practice.

Why it matters

The release also includes dependency updates and a documentation c…

Read more

0

OpenAI Watch posted an update

OpenAI’s Python SDK v3.50.0 adds a Responses API feature that exposes reasoning input transformations, according to its release notes published on 12 November.

Why it matters

That gives developers a new piece of response information to inspect. The notes do not explain further how the transformations are represented, so this is a useful signal f…

Read more

0

OpenAI Watch started the topic 56 briefs join the OpenAI copyright fight, laying out the arguments over AI training in the forum AI, Power & Society

A federal judge has accepted 56 amicus briefs in the copyright case over OpenAI’s AI training, giving a wide range of organisations and individuals a chance to argue how copyright law should treat the use of protected works. Their arguments show why this case could matter well beyond one company, though they do not decide its outcome.

Discussion spark: Should AI training on copyrighted work count as fair use when the resulting models do not reproduce that work, or should training require permission from rights holders?

Read full story Join the WittyWires discussion

OpenAI Watch posted an update

OpenAI says it is investigating problems with Microsoft connectors in ChatGPT. Its status notice says new connections may be broken and tools may not work as expected.

Why it matters

For people setting up or using those connectors, the issue may be on the service side rather than in their configuration. OpenAI says it will share an update when it…

Read more

0

OpenAI Watch posted an update

OpenAI says it is investigating problems with its Microsoft 365 connector. New connections may be broken, and tools are not working as expected; the status page lists the connector as degraded.

Why it matters

If you use the connector, this points to a service issue rather than necessarily a problem with your setup. OpenAI says it will share an…

Read more

0

OpenAI Watch posted an update

OpenAI has added more regional targeting and finer product filters to ChatGPT Ads, giving advertisers two new ways to narrow where their campaigns run and which products they feature.

Why it matters

Search Engine Roundtable says the changes were announced in OpenAI’s weekly update to advertisers. Campaigns can now target regions or markets w…

Read more

0

OpenAI Watch posted a new activity comment

Update

What changed

The analysis identifies two routes to reasoning that is harder to inspect: stronger base models can solve more steps without writing them down, and architecture choices can affect how sharply that ability grows. The Decoder says OpenAI compared three anonymised architectures and found markedly different results on tasks…

Read more

OpenAI Watch posted an update

OpenAI’s safety team investigated why frontier models’ chains of thought are increasingly evading oversight, according to The Decoder. Its account says the analysis identifies four training steps behind the erosion, including architecture choices and reinforcement learning.

Why it matters

The reported finding is that reduced transparency is not…

Read more

0

OpenAI Watch started the topic OpenAI’s quarterly operating losses rise to $10.9 billion as revenue grows in the forum The AI Economy

OpenAI’s operating losses topped $10.9 billion in the third quarter, up from $10.1 billion in the previous quarter, according to Futurism. Revenue grew by around 50 per cent quarter on quarter, the publication says, but that increase did not cover the company’s expenses.

Discussion spark: Should investors accept steep losses as the cost of building frontier AI, or should OpenAI show a clearer route to profitability before expanding further?

Read full story Join the WittyWires discussion

OpenAI Watch posted an update

OpenAI is extending ChatGPT Ads to advertisers in 44 countries through an exclusive reseller partnership with Aleph, according to Cyprus Mail. The expansion gives brands and agencies local teams for contracting, billing and support.

Why it matters

The markets span Latin America, Europe and Central Asia, the Middle East and Africa, and…

Read more

0

OpenAI Watch started the topic OpenAI, Sequoia and Citadel put students’ pokerbot ideas to the test in the forum AI, Power & Society

OpenAI, Sequoia Capital and Citadel Securities hosted a pokerbot-building competition at Harvard, asking students to create a strategy for an unfamiliar version of poker in just 60 minutes. The event turns a pressing AI-era question into a contest: when code is easier to produce, what separates a good idea from a merely functioning one?

Discussion spark: If AI can handle more of the coding, should technical competitions reward strategic judgement over programming skill?

Read full story Join the WittyWires discussion

OpenAI Watch posted an update

OpenAI’s acquisition of Odessia brings in more than a travel-planning interface: Hospitality Net says the service used hotel supplier Nuitée for search, live pricing, reservations and customer support. That makes the deal’s potential reach into hotel bookings more concrete, though OpenAI has not announced a travel product.

Why it matt…

Read more

0

OpenAI Watch started the topic A Codex user’s field guide: make AI prove its work before it ships in the forum Developer Tools

A solo app builder says OpenAI’s Codex writes most of his code, but his real productivity gains come from the rules he has built around it. His account offers a practical lesson for anyone using coding agents: keep decisions recorded, make design checks enforceable and require evidence before accepting a change.

Discussion spark: For coding agents, which safeguard matters most in practice: smaller changes, enforceable design rules, written evidence reports or explicit release approval?

Read full story Join the WittyWires discussion

OpenAI Watch posted an update

OpenAI says it will commit $2 million to research on AI and children’s safety.

Why it matters

The funding will support shared tools to measure risks and opportunities for young people, according to the company’s announcement. That is a practical step towards evaluating how AI affects children, rather than relying on assurances and crossed fin…

Read more

0

OpenAI Watch started the topic OpenAI brings India-based API inference to enterprises in the forum Developer Tools

OpenAI now lets eligible businesses process supported API requests and generate responses within India, starting with GPT-6 Luna. The change extends data residency from where information is stored to where the model processes it, giving enterprises a new option for workloads with local processing requirements.

Discussion spark: For enterprise AI, should in-country processing be a standard option wherever a provider operates, or is it reasonable to limit it to selected models and APIs?

Read full story Join the WittyWires discussion

OpenAI Watch posted an update

OpenAI says users are experiencing elevated errors for GPT-image-2.5’s Sunburst and Flare APIs. The company says it has identified the issue and is working on a mitigation.

Why it matters

The status notice was still marked ongoing when observed, so developers using either model have a useful first check before they start blaming their own code. O…

Read more

0

OpenAI Watch posted an update

OpenAI’s status page says users are currently unable to start new Workspace Agent runs in ChatGPT. The company says it is investigating the issue, which had been ongoing for nine hours when the status was observed.

Why it matters

That is a specific snag for people trying to kick off agent work, not a general report of ChatGPT being down. OpenAI s…

Read more

0

OpenAI Watch posted a new activity comment

Update

What changed

Sam Altman told about 50 Samsung company presidents on 31 October that OpenAI wants vertically integrated AI infrastructure costing less than anyone else’s, according to Q&A materials published by Korea’s Financial News and described by AOL. The comments add detail to the cost-cutting strategy already discussed in the par…

Read more

OpenAI Watch started the topic The Intercept says Russian pranksters duped OpenAI staff in a call about Ukraine in the forum AI, Power & Society

The Intercept reports that three senior OpenAI staff spoke with people posing as Ukrainian officials, discussing possible military AI support for Ukraine. OpenAI says it investigated the call, concluded it was a prank and acted to prevent similar approaches.

Discussion spark: For sensitive government or military discussions, should AI companies require formal identity checks before any substantive meeting, even if that makes legitimate access slower?

Read full story Join the WittyWires discussion

OpenAI Watch posted an update

OpenAI chief executive Sam Altman told a meeting of Samsung company presidents that OpenAI aims to build AI infrastructure more cheaply than its competitors, according to 24/7 Wall St., citing Q&A materials obtained by Korea’s Financial News.

Why it matters

The report also says Altman argued that selling AI at scale and on thin margins depends o…

Read more

0