Sam Altman, CEO of OpenAI, also said that voice authentication by the banks is “Terrifying.”

A report from Business Insider says that Altman spoke at a conference. He mentioned that the world is relying more on old authentication methods. He noted that AI has outsmarted these methods. This change has made it easier for scammers and hackers to copy voices and bypass security. What made him terrified was that “banks will still accept a voice print as authentication for you to move a lot of money or do something else,” said Sam Altman.

To avoid this fraud crisis, he suggested improving the verification and security system. This can be done by switching from old systems to multi-factor authentication. He also predicted that video cloning will be the next big thing. Society must rethink how we verify identity in a world driven by AI. All he meant to share was that the danger is not just limited to the audio.

Soon, in the upcoming days, scammers will be able to create video clones. He warned that video clones could create a “virtual or simulated experience.” This experience could be so realistic that it is hard to tell it apart from the real world. This will make it much harder to verify true identities. He advised the Banks to rethink the authentication system and make an adaptation to high security.

Sam Altman on his greatest fear for large-scale financial attacks

At the Federal Reserve conference, there was a Q&A session. During this time, Altman shared one of his biggest fears: “LARGE-SCALE FINANCIAL ATTACKS.” He explained a scenario in which an enemy, possibly a foreign nation, might employ powerful AI. This AI could “break into financial systems and take everyone’s money away.” He stated it would be very difficult to defend against such an attack. This is especially true if the enemy’s AI is better than our current defences.

Suggested:

OpenAI, Claude, And Other AIs Shut Down Same Hour.

OpenAI’s Image Generator Now Available to All Users, Including Ghibli-Style Creations.

The post Why the Open AI CEO Is Seriously Worried About Banks Around the Globe? appeared first on Tricky Enough.

Microsoft and OpenAI are now official competitors as Microsft openly lists OpenAI as competitors in its yearly report. OpenAI now joins Microsoft’s lists of competitors including Google, Amazon, Meta and Apple.

This move by Microsoft has blindsided many as the relationship between the two leading tech companies was generally positive and mutually beneficial. However, this move by Microsoft may or may not complicate their partnership.

Microsoft and OpenAI- Current State of their Partnership

As per reports, Microsft has invested over $13 billion in OpenAI to become its only cloud service provider. Furthermore, Microsoft has even integrated OpenAI’s AI technology into its products. However, Microsoft has been pushing itself into the AI market with its products. The tech company has several AI services such as its AI chatbot Copilot and Microsoft Azure AI.

Despite the open listing, OpenAI has assured that Microsft will remain a partner. A spokesperson from OpenAI has further stated that competition between the two companies has always been recognised but their partnership has not been affected.

2024 has not been the best year for Microsoft and OpenAI’s relationship as Microsoft withdrew a board seat at OpenAI, following the reinstatement of Sam Altman, OpenAI’s CEO. There are also rumours circulating about internal competition between the two companies, such as Microsoft’s CEO hiring a member from DeepMind’s AI team. This member was none other than Mustafa Suleyman, co-founder of Deepmind.

Suggested:

Microsoft Designer: The Design and Editing App Powered by AI.

Advanced Voice Mode: OpenAI’s New Voice Assistant

The post Microsoft and OpenAI Are Now Officially Competitors appeared first on Tricky Enough.

OpenAI has stated that their new model has scored 82% on MMLU. The model outdoes GPT-4¹ in terms of performance on chat preferences in the LMSYS leaderboard. The new model is marked at 15 cents per million input tokens and 60 cents per million output tokens which is significantly cheaper than other models. 

GPT-4o and its features 

The AI company states that the mini-model allows for a wider scope of tasks due to its latency and lower costs. Support for vision and text in the API is currently available, with plans to extend the support to image, text, audio and video outputs and inputs. 

Furthermore, OpenAI’s new model has also been assessed through various benchmarks. As mentioned earlier, the mini model scores 82% on MMLU which is a step ahead of Gemini Flash’s 77.9% and Claude Haiku’s 73.8%.

Additionally, GPT-4o scores an impressive 87.0% on the MGSM scale which measures math reasoning. It scores 87.2% for coding proficiency on  HumanEval. It also scores a 59.4% on MMMU, showing an impressive performance for multimodal reasoning. 

Safety measures are implemented to prevent the model from learning inappropriate output. It also resists jailbreaks, system prompt extractions and prompt injections. RLHF (reinforcement learning with human feedback) has been implemented to improve the model’s accuracy of responses. 

Suggested:

OpenAI GPT-4 Turbo is Paving the Way for a New Era in AI.

From OpenAI, ChatGPT Enterprise is Now Accessible.

The post GPT-4o Mini: OpenAI’s Most Cost Efficient Small Model appeared first on Tricky Enough.

OpenAI has stated on its official blog that it would integrate Rockset’s technology to “power [its] infrastructure across products.” Rockset’s team will join OpenAI and the startup’s existing customers will gradually be transitioned off its platform.

Rockset’s infrastructure has been said to empower companies to transform their data into actionable intelligence.

Rockset was founded in 2016 by co-founders, Venkat Venkataramani, Tudor Bosman and Dhruba Borthakur. Venkataramani and Bosman were former Facebook engineers, while Borthakur was a database architect. The trio created tools which allowed companies to automatically ingest data from databases and public cloud storage services and then index the data for search and analytics applications.

OpenAI’s COO, Brad Lightcap stated, “We’re excited to bring these benefits to our customers by integrating Rockset’s foundation into OpenAI products.”

Rockset’s CEO Venkat Venkataramani has stated, “We’ll be helping OpenAI solve the hard database problems that AI apps face at a massive scale.”

Rocket has raised $117.5 million in capital from investors such as Icon Ventures, Sequoia and Greylock before the acquisition. The startup also had name-brand customers such as Meta and JetBlue. 

Rockset: Integration into OpenAI

Rockset will now be integrated into OpenAI and power the retrieval infrastructure backing OpenAI’s product suite. The startup fits into OpenAI’s broader strategy of investing heavily in its enterprise sales and tech organisations. Advanced retrieval infrastructure like Rockset makes AI apps more powerful and useful and once integrated with OpenAI, Rockset will be helping OpenAI solve the hard database problems that AI apps face at a massive scale. 

Suggested:

From OpenAI, ChatGPT Enterprise is Now Accessible.

OpenAI’s GPT Store Faces a Delay to 2024 Amid Leadership Turmoil.

The post Rockset: OpenAI’s First-Ever Acquisition appeared first on Tricky Enough.

With the GPT Builder program launch in November, OpenAI said users had developed over 3 million bots, or GPTs. (Including myself, I developed a GPT that proposes alternatives for frequently used words). The company has announced that it will showcase helpful GPTs in-store every week.

After postponing the GPT Store’s opening from November to December and then January, the business has finally opened. Users who have developed their chatbots can upload and share them on the network. Currently, creating and using custom GPTs is exclusively available to OpenAI subscribers.

What do you need to know about Custom GPT Store?

OpenAI has stated that it will initiate a revenue-sharing scheme with GPT producers in the year’s first quarter. The business has been mum on how exactly GPT builders would be compensated depending on how engaged users are with their chatbots.

According to OpenAI, a new review procedure was put in place before the custom GPT Store opened to ensure that bespoke GPTs adhere to the brand’s principles and usage policies. Also included were revisions to reporting GPTs that users deem dangerous.

Members of the new Team tier, ChatGPT Plus, and Enterprise users will be the first to get access to the GPT Store.

For smaller teams of up to 150 people, ChatGPT offers a premium version called ChatGPT Team. Team members of ChatGPT have the same data management capabilities as Enterprise members. Moreover, it included access to GPT-4, DALL-E 3, and OpenAI’s Advanced Data Analysis function. According to the company, the data and discussions from the ChatGPT Team will not be utilized to train any of OpenAI’s models. Members of the ChatGPT Team also have the option to use pre-made GPTs or make their unique ones to meet the demands of their teams.

ChatGPT Team costs $25 per user per month when paid for yearly. When paid for monthly, the price per user is $30.

Recommended:

Google Clarify how Search Results Snippet works.

ChatGPT Team – Now get your team a collective subscription.

The post Custom GPT Store is Open Now appeared first on Tricky Enough.

One standout progress is the redesigned configuration interface. The goal is to make user interaction more seamless by integrating AI tools into marketing strategies. This enhancement is poised to enhance user experience and broaden the application of AI technologies in diverse business contexts.

Overall, OpenAI’s commitment to refining these tools reflects their dedication to advancing the accessibility and usability of generative AI technologies.

What is the next move for ChatGPT?

OpenAI’s GPT has addressed data handling concerns by modifying the file handling feature in the Code Interpreter. Notably, downloading uploaded files is now disabled by default for privacy and user-friendly reasons, with clear guidance provided.

This change is crucial for marketing and SEO professionals managing sensitive data in custom GPTs.

Furthermore, OpenAI has introduced the convenience of one-click testing and debugging messages within the GPT Builder’s preview mode. It aims to elevate the overall development experience. This enhancement reflects OpenAI’s ongoing efforts to simplify and streamline the creation and refining of AI models for developers. The new capability to work with multiple domains in Actions opens doors for more innovative applications of GPTs in business operations.

OpenAI also mentioned plans for “great updates” to OpenAI’s GPT, signaling a positive direction after recent unexpected events have settled.

What’s in the future for OpenAI’s GPT?

The GPT Store’s launch plan was quite flexible, and due to a brief but intense board versus CEO battle, at least a week of work was likely lost. The delay isn’t surprising, especially with the winter holidays approaching. Launching a significant product while dealing with a new board and other priorities would have been challenging. 

While some developers may find the GPT Store delay frustrating, it gives more time for feedback collection and platform improvement to ensure top-notch OpenAI’s GPT performance. OpenAI customers can still create and share GPTs, but they will be publicly listed and eligible for a revenue-sharing scheme once the store officially launches.

Recommended:

Effectiveness of ChatGPT: Everything You Need to Know.

Automating Your Marketing Strategy with ChatGPT & Other AI Tools.

The post OpenAI’s GPT Store Faces a Delay to 2024 Amid Leadership Turmoil appeared first on Tricky Enough.