Your Followed Topics

Top 32 openai News Today

#1
AI outperforms doctors in Harvard trial of emergency triage diagnoses
#1 out of 3211.11%

AI outperforms doctors in Harvard trial of emergency triage diagnoses

https://www.theguardian.com/technology/2026/apr/30/ai-outperforms-doctors-in-harvard-trial-of-emergency-triage-diagnoseshttps://mashable.com/article/can-ai-replace-a-doctorhttps://www.npr.org/2026/04/30/nx-s1-5804474/ai-doctors-openai-patient-care-diagnosis
Theguardian.com and 6 more
  • The AI advantage was pronounced in fast triage decisions with minimal information, suggesting AI can aid rapid ER judgments.
  • In real-world ER data, AI achieved 67% exact/near diagnoses at triage and 81% at admission, surpassing two expert doctors.
  • Experts say AI should augment clinicians as a second opinion and workflow support in the ER, not replace doctors.
Vote 18
0
#2
The craziest part of Musk v. Altman happened while the jury was out of the room
#2 out of 32
technology18m ago

The craziest part of Musk v. Altman happened while the jury was out of the room

  • The court heard a $97.4 billion bid for OpenAI assets connected to xAI, discussed by Musk and associates.
  • The judge questioned the accuracy and foundation of the bid figures during direct examination.
  • Birchall claimed he did not know whether he discussed the bid with Elon Musk or Shivon Zilis.
  • The jury was dismissed early as the court debated objections to Birchall’s testimony.
  • Judge Gonzalez Rogers pressed for clarity on whether the xAI bid involved privileged communications.
  • Birchall explained the bid’s purpose was to account for nonprofit asset value amid restructuring.
  • The proceedings included on-the-record exchanges with OpenAI’s lawyers about testimony admissibility.
  • The judge asked who passed the note prompting the discussion of the bid during testimony.
  • The coverage notes corrections to names and dates, highlighting live updates from Elizabeth Lopatto.
  • The coverage underscores the ongoing tension between business deals and legal discovery in tech trials.
Vote 0
0
#3
How Shivon Zilis Operated as Elon Musk’s OpenAI Insider
#3 out of 3220.45%

How Shivon Zilis Operated as Elon Musk’s OpenAI Insider

  • Musk v. Altman trial centers on Shivon Zilis, OpenAI adviser and liaison since 2016.
  • Zilis joined OpenAI in 2016 as an adviser and later served on its nonprofit board.
  • Musk described Zilis as a 'chief of staff' or 'close adviser' in court testimony.
  • Court records show Zilis served as a liaison, keeping OpenAI connected with Musk after 2018.
  • Emails show Zilis coordinated discussions on OpenAI’s governance and equity sharing.
  • Zilis advised on fundraising progress and OpenAI initiatives, including an AI game project.
  • Zilis balanced roles at Neuralink, Tesla, and OpenAI, coordinating multiple priorities.
  • Text messages show strategic advice between Zilis and Altman after funding rounds.
  • Zilis advised on how to respond to Musk’s concerns about OpenAI’s mission trajectory.
  • The WIRED profile links Zilis to leadership decisions shaping OpenAI during its early years.
  • OpenAI’s management and structure discussions are a throughline of Zilis’s work.
  • The article situates Zilis as a central but discreet figure in OpenAI’s early ecosystem.
Vote 0
0
#4
4 things you missed from Day 4 of the Musk v. Altman trial, including testimony of Musk's money manager
#4 out of 3214.29%

4 things you missed from Day 4 of the Musk v. Altman trial, including testimony of Musk's money manager

  • Musk described xAI as very small, about a tenth the size of OpenAI, positioning it as the smallest among his AI ventures.
  • Judge Yvonne Gonzalez Rogers barred discussion of an AI extinction scenario during questions to Musk.
  • Birchall testified Musk donated about $38 million to OpenAI and made roughly 60 donations in total.
  • Birchall described a 2018 office-sharing arrangement involving Neuralink related to OpenAI's security costs.
  • Upcoming testimony will include Greg Brockman, OpenAI’s president, cited for diary entries about the pre-2018 period.
  • The trial centers on Musk's challenge to OpenAI's nonprofit-to-profit transition and competition with rivals.
  • Birchall noted OpenAI’s 2018 office-sharing arrangement with Neuralink involved security concerns.
  • The article highlights that Brockman’s diary entries are contested by OpenAI as cherry-picked evidence.
  • The trial will continue with further witness testimony from OpenAI backers and executives next week.
  • Musk reiterated his stance that xAI is not a threat and ranked rivals by size.
Vote 0
0
#5
OpenAI to Launch Codex Mobile App for iPhone as a General Productivity Tool
#5 out of 3250.00%

OpenAI to Launch Codex Mobile App for iPhone as a General Productivity Tool

  • OpenAI plans a dedicated iPhone app for Codex to expand its use beyond coding.
  • Codex would become a general productivity assistant, not limited to programmers.
  • The app aims to help with writing, planning, and research in daily life.
  • OpenAI positions Codex as part of a broader move to practical AI in everyday life.
  • Codex on mobile reflects OpenAI’s strategy to broaden access to AI tools.
  • The move accompanies broader shifts in OpenAI’s partnership and strategy with Microsoft.
  • Codex was previously viewed mainly as a coding and developer tool.
  • The planned mobile app would serve as the main interaction point for Codex features.
  • Codex’s expansion aligns with OpenAI’s broader goal of practical, easy-to-use AI.
  • The development is framed as bringing Codex to everyday tasks like writing and planning.
Vote 0
0
#6
ChatGPT now asks you how hard it should think before you even type, just like Gemini does — and it completely changes how you use it
#6 out of 32
technology14h ago

ChatGPT now asks you how hard it should think before you even type, just like Gemini does — and it completely changes how you use it

  • OpenAI moved the model picker into the ChatGPT prompt area for Plus, Pro, and Business plans.
  • Instant uses GPT-5.3 and is best for quick answers, while Thinking targets harder tasks with extended reasoning.
  • Configure lets users specify exact model versions back to GPT-5.4 and older.
  • The article notes a hidden thinking-time setting with Standard and Extended options.
  • Pro remains the highest-capability option for the hardest tasks on the platform.
  • Auto-select can reduce manual changes but raises ethical considerations about resource use.
  • The update aligns ChatGPT with Gemini’s interface approach by exposing model choices more clearly.
  • The piece outlines when to use Instant versus Thinking for different tasks.
  • The article summarizes background on the model options’ prior locations and expectations for future rollouts.
  • The story provides context on model version availability and ecosystem implications.
  • The article emphasizes that interface changes could affect the pace of model usage among users.
Vote 0
0
#7
OpenAI’s subtle drift from Microsoft has become an aggressive move toward Amazon
#7 out of 32
technology17h ago

OpenAI’s subtle drift from Microsoft has become an aggressive move toward Amazon

  • OpenAI expands its cloud partnership with Amazon, bringing its models to AWS.
  • OpenAI restructures its relationship with Microsoft, ending exclusive licensing and revenue shares.
  • Analysts say the moves signal a broader shift in OpenAI's cloud strategy amid tight compute capacity.
  • OpenAI insists the Microsoft and Amazon deals are not related, according to Denise Dresser.
  • Amazon's $50 billion investment in OpenAI supports expanded AWS training and services.
  • OpenAI and Amazon will jointly develop customized models for Amazon's engineering teams.
  • Microsoft previously relied on OpenAI but has sought diversification away from exclusive ties.
  • UBS analysts noted Microsoft made more concessions than gains in the new terms.
  • OpenAI emphasizes meeting customers where they are to deliver the best models in the best environments.
  • Microsoft's Copilot and Anthropic partnerships illustrate continued diversification.
  • OpenAI's revenue growth and IPO timing remain under scrutiny as partnerships evolve.
Vote 0
0
#8
Gabe Newell was an enthusiastic supporter of OpenAI in 2018, donating $20 million and even acting as the sole member of an 'informal advisory board'
#8 out of 32

Gabe Newell was an enthusiastic supporter of OpenAI in 2018, donating $20 million and even acting as the sole member of an 'informal advisory board'

  • Gabe Newell donated $20,008,279 to OpenAI in 2018, as part of early funding
  • Newell served as an informal advisor to OpenAI alongside Elon Musk in the early days
  • OpenAI listed Newell’s donation on its website in 2018, confirming early backing
  • The court filings show Newell’s continued involvement with tech ventures beyond OpenAI
  • The disclosures connect Newell’s interests with neuroscience and gaming leadership
  • The Musk v. Altman case reveals early OpenAI funding and advisory networks
  • Newell’s involvement underscores gaming leaders' ties to AI research funding
  • The reporting highlights how OpenAI benefited from early donations during its formative period
  • The court filings describe Newell’s role alongside other OpenAI backers in its early rounds
Vote 0
0
#9
OpenAI Rolls Out ‘Advanced’ Security Mode for At-Risk Accounts
#9 out of 32
technology10h ago

OpenAI Rolls Out ‘Advanced’ Security Mode for At-Risk Accounts

  • OpenAI rolled out Advanced Account Security, a new optional protection for at-risk ChatGPT and Codex accounts.
  • Users must use two physical security keys or passkeys, eliminating regular passwords for protected accounts.
  • Recovery options shift from email/SMS to recovery keys or backup passkeys.
  • OpenAI partners with Yubico to offer lower-cost YubiKey bundles for users.
  • Support for account recovery is removed; attackers cannot target support portals to regain access.
  • Security alerts will notify users of any login to the locked-down account.
  • The feature targets high-stakes users, including journalists and researchers.
  • Trusted Access program members must enable the feature or attest to phishing-resistant authentication.
  • Advanced Account Security is framed as part of a broader cybersecurity strategy.
Vote 0
0
#10
Elon Musk Seemingly Admits xAI Has Used OpenAI's Models to Train Its Own
#10 out of 32
technology10h ago

Elon Musk Seemingly Admits xAI Has Used OpenAI's Models to Train Its Own

  • Musk testified that distillation is common and that xAI used other AIs to validate its models, signaling cross-industry practices.
  • The trial referenced OpenAI’s efforts to protect its models against distillation, reflecting ongoing regulatory and competitive concerns.
  • Analysts note that distillation helps smaller models imitate larger ones, making them cheaper and faster to run.
  • The testimony occurred amid broader questions about OpenAI’s governance and the future of AI innovation in a competitive landscape.
  • The article places the exchange in the context of the Musk v. Altman trial, where jurors weigh OpenAI’s evolution and OpenAI’s original mission.
  • The coverage mentions other industry dynamics, including responses from Anthropic and policy makers.
  • OpenAI’s memo to a House committee highlighted a focus on preventing foreign distillation by rivals.
  • The coverage underscores that distillation is used to validate AI across multiple labs, not only OpenAI and xAI.
  • The article cites broader policy signals from government officials about global AI competition and safeguarding innovation.
Vote 0
0

Explore Your Interests

Unlimited Access
Personalized Feed
Full Experience
or
By continuing, you agree to the Privacy Policy.. You also agree to receive our newsletters, you can opt-out any time.

Explore Your Interests

Create an account and enjoy content that interests you with your personalized feed

Unlimited Access
Personalized Feed
Full Experience
or
By continuing, you agree to the Privacy Policy.. You also agree to receive our newsletters, you can opt-out any time.

Advertisement

Advertisement