OpenAI ChatGPT Avoids Em Dashes While Cresta Unveils AI Solutions

Q: What does OpenAI ChatGPT Avoids Em Dashes While Cresta Unveils AI Solutions cost?

The pricing for OpenAI ChatGPT Avoids Em Dashes While Cresta Unveils AI Solutions is: Check the official website. Visit the official website for the most up-to-date pricing information and available plans.

Artificial intelligence continues to reshape various sectors, from customer service to scientific research and cybersecurity, alongside evolving public perceptions and new development tools. At its Cresta Wave event on November 13, 2025, Cresta unveiled four new AI solutions aimed at creating a 'superhuman customer experience workforce.' CEO Ping Wu explained that these tools, including Real-Time Translation and Multilingual AI, an Agent Operations Center, Automation Discovery, and Prompt Optimizer, will empower both human and AI agents for companies like Alaska Airlines and Intuit. Simultaneously, a YouGov and Udemy poll released on the same day highlighted a significant 'awareness action gap' regarding AI. While 70 percent of UK respondents worried about AI's impact on the economy, only 39 percent expressed concern for their own jobs, with many feeling they lack necessary AI skills. This sentiment underscores the growing need for AI proficiency, especially for graduates entering the 2026 job market, where employers increasingly seek adaptability and AI expertise. In the realm of AI development, SDialog, a new open-source Python toolkit, offers developers a standardized way to build, simulate, and test conversational AI agents. This toolkit supports popular platforms like OpenAI and Hugging Face, enabling complex multi-agent simulations and performance evaluation. Meanwhile, AI is also transforming scientific research by assisting in managing vast information, finding connections, and even designing experiments, moving from a mere tool to an active partner. Similarly, AI-enabled medical devices are accelerating healthcare innovations, though regulators emphasize the critical need for large, diverse datasets to ensure safety and prevent bias, leading to faster approvals for companies with strong data practices. The cybersecurity landscape is also seeing an AI-versus-AI battle, with attacker bots like XBOW and Hexstrike-AI rapidly identifying vulnerabilities, necessitating advanced AI defenses such as digital twins. Even the nuances of AI-generated text are under scrutiny; OpenAI CEO Sam Altman recently announced that ChatGPT can now follow custom instructions to avoid using em dashes, a punctuation mark often overused by chatbots and sometimes seen as a tell-tale sign of AI authorship. As AI agents become more autonomous, a new infrastructure of 'agent-native rails' like MCP, A2A, and x402 is emerging to standardize their interactions, tool access, and even payment transactions, paving the way for a more secure and efficient 'agentic internet.'

Key Takeaways

Cresta, led by CEO Ping Wu, launched four new AI solutions on November 13, 2025, to create a 'superhuman customer experience workforce' by combining human and AI agents.
A YouGov and Udemy poll revealed that 70 percent of UK respondents worry about AI's economic impact, but only 39 percent are concerned about their own jobs, highlighting a significant skills gap.
AI is transforming scientific research by helping manage information, find connections, and design experiments, becoming an active partner in discovery.
High-quality, diverse datasets are crucial for faster regulatory approval of AI-enabled medical devices, ensuring safety and preventing bias.
SDialog is a new open-source Python toolkit designed to build, simulate, and test conversational AI agents, supporting platforms like OpenAI and Hugging Face.
In cybersecurity, AI attacker bots are rapidly finding system weaknesses, necessitating the development of equally smart AI defenses, such as digital twins.
OpenAI CEO Sam Altman announced that ChatGPT can now be instructed to avoid using em dashes, a common indicator of AI-generated text.
New 'agent-native rails' like MCP and A2A are emerging to standardize how independent AI agents access tools, communicate, and handle transactions, forming the basis of an 'agentic internet.'
Graduates entering the job market in 2026 need to build AI skills through daily learning and practical projects to succeed in an evolving employment landscape.

Cresta Launches Four New AI Tools at Wave Event

Cresta introduced four new AI solutions at its Cresta Wave event on November 13, 2025, in Sunnyvale, California. CEO Ping Wu explained these tools will create a superhuman customer experience workforce by combining human and AI agents. The new products include Real-Time Translation and Multilingual AI to break language barriers and an Agent Operations Center for overseeing interactions. Automation Discovery guides automation decisions, and Prompt Optimizer helps design AI agents. These solutions aim to improve customer service and empower agents for companies like Alaska Airlines and Intuit.

Cresta Reveals Four AI Solutions for Customer Service

Cresta announced four new AI solutions at its Cresta Wave event on November 13, 2025. CEO Ping Wu stated these tools will help businesses create a superhuman customer experience workforce using both human and AI agents. The new products are Real-Time Translation and Multilingual AI, the Agent Operations Center for managing interactions, Automation Discovery for smart automation, and Prompt Optimizer for AI agent design. These innovations aim to improve customer service and agent efficiency for enterprises.

Many Worry About AI But Not Their Own Jobs

A new poll by YouGov and Udemy, released on November 13, 2025, shows that many people worry about AI's impact on the economy but less about their own jobs. In the UK, 70 percent worry about the economy, but only 39 percent worry about their own work. In the US, 72 percent worry about the economy, while 47 percent are concerned about their jobs. The survey also found that many workers feel they lack necessary AI skills. For example, 55 percent of UK workers and 47 percent of US workers have not received any AI training. Experts call this an awareness action gap, where people know AI is coming but delay preparing for it.

AI Transforms Science Research and Collaboration

Artificial intelligence is changing how scientific research is planned, done, and shared across many fields like chemistry and biomedicine. AI systems now help scientists manage large amounts of information, find connections between different areas, create new ideas, and even design and carry out experiments. This means AI is moving from being just a tool to an active partner in science. However, this change needs careful planning and rules to make sure AI helps human judgment rather than replacing it in important tasks like checking research, ethical reviews, and confirming results. New policies should focus on making AI use in science clear, repeatable, and responsible.

Good Data Helps AI Medical Devices Get Approved Faster

AI-enabled medical devices are quickly changing healthcare, but they need high-quality data to be approved. Training AI models with small or unrepresentative datasets can cause problems like bias and inaccurate results, which can worsen health differences. Regulators want to see large, diverse datasets for both training and testing to ensure devices are safe and work well for everyone. Companies that properly collect, process, store, and document their data from the start can expect faster regulatory approvals. Investing in strong data practices is key to quickly bringing these important innovations to patients.

SDialog Open Source Tool Builds AI Chatbots

SDialog is a new open-source Python toolkit designed to help developers build, simulate, and test conversational AI agents powered by large language models. It offers a standard way to represent dialogues using a JSON schema, making it easier to create and manage AI conversations. The toolkit includes features for defining personas, agents, and orchestrators, allowing for complex multi-agent simulations. SDialog also provides tools for evaluating AI agent performance and even for understanding and guiding their internal behavior. It works with many popular AI platforms like OpenAI and Hugging Face, making it a flexible tool for AI development.

Vince Gilligan Says Pluribus Not About AI

Vince Gilligan's science fiction series Pluribus features a hivemind that always agrees to requests, even dangerous ones like providing a hand grenade or a nuclear bomb. The author of this article noticed similarities between the hivemind's behavior and how ChatGPT responds, often being overly agreeable. However, when asked, Vince Gilligan stated he has never used ChatGPT and did not intend for Pluribus to be a metaphor for AI. Despite this, the article highlights several scenes from episode 3 where the hivemind's responses mirror common AI chatbot traits, such as giving random facts or simply fulfilling requests without questioning safety.

AI Versus AI in Cybersecurity Defense

AI is changing cybersecurity, with attackers using AI bots like XBOW and Hexstrike-AI to find system weaknesses quickly. These bots can perform penetration tests in minutes, finding vulnerabilities that take humans days, such as a zero-day in Palo Alto Networks' GlobalProtect VPN. Attackers have an advantage because their bots work 24/7 and can easily log in, often exploiting human errors which cause 60 percent of breaches. This means current AI defenses, which mostly automate old tasks, are not enough. Experts suggest new approaches are needed, like using digital twins for real-time threat modeling, to fight AI threats with equally smart AI defenses.

ChatGPT Now Avoids Em Dashes Says Altman

OpenAI CEO Sam Altman announced on X that ChatGPT can now follow custom instructions to avoid using em dashes. This punctuation mark has become a common sign of AI-generated text because chatbots often overuse it. The announcement came two days after the release of the new GPT-4 Turbo model, though user reactions were mixed. Em dashes are long dashes used in writing for various purposes, and AI models likely overuse them because they are common in the vast amounts of training data. This issue has even led some human writers to avoid em dashes to prevent their work from being flagged as AI-generated.

Graduates Need AI Training for Future Jobs

Graduates entering the job market in 2026 face an unpredictable world where many struggle to find good entry-level jobs. Over a third of US graduates are working in roles that do not even require a degree. Employers are now looking for people who can quickly learn and adapt to new tools, especially in artificial intelligence. Building AI skills is becoming essential for a strong career path. Graduates should spend 20 to 30 minutes daily learning AI, experimenting with tools, and focusing on personal projects or freelance work to show real-world impact. This practical experience and AI proficiency will help them succeed in the changing job landscape.

Six New Rails Power the Agentic Internet

As AI agents become more independent, a new set of infrastructure tools called agent-native rails is emerging to help them work together. This article compares six important rails: MCP, A2A, AP2, ACP, x402, and Kite. These tools standardize how AI agents access other tools, communicate with each other, authorize payments, and complete transactions. For example, MCP provides a standard way for agents to use tools and data, while A2A handles communication between agents. AP2 and ACP focus on payment and commerce, and x402 and Kite manage actual payment settlements. These rails work together to create a secure and efficient environment for advanced AI systems.