This year's hot new tool for chefs? ChatGPT.
If all goes according to plan, he will keep prompting the program to refine one of Jill's recipes, along with those of eight other imaginary chefs, for a menu almost entirely composed by artificial intelligence.
Get Winter Soup Club
A six-week series featuring soup recipes and cozy vibes, plus side dishes and toppings, to get us all through the winter.
Enter Email
Sign Up
'I want it to do as much as possible, short of actually preparing it,' Achatz said.
Advertisement
As generative AI has grown more powerful and fluent over the past decade, many restaurants have adopted it for tracking inventory, scheduling shifts, and other operational tasks. Chefs have not been anywhere near as quick to ask the bots' help in dreaming up fresh ideas, even as visual artists, musicians, writers, and other creative types have been busily collaborating with the technology.
That is slowly changing, though. Few have plunged headfirst into the pool in quite the way Achatz is doing with his menu for Next, but some of his peers are also dipping exploratory toes into the water, asking generative AI to suggest spices, come up with images showing how a redesigned space or new dish might look, or give them crash courses on the finer points of fermentation.
Advertisement
'I'm still learning how to maximize it,' said Aaron Tekulve, who finds the technology helpful for keeping track of the brief seasonal windows of the foraged plants and wild seafood from the Pacific Northwest that he cooks with at Surrell, his restaurant in Seattle. 'There's one chef I know who uses it quite a bit, but for the most part I think my colleagues don't really use it as much as they should.'
Goat sausage with butter beans and focaccia croutons at Houseman in Manhattan, May 29, 2025. Ned Baldwin, the restaurant's chef and owner, asked for ChatGPT's help in understanding the technical details of sausage-making.
EMON HASSAN/NYT
The pinball-arcade pace of a popular restaurant can make it hard for chefs to break with old habits. Others have objections that are philosophical or aesthetic.
'Cooking remains, at its core, a human experience,' chef Dominique Crenn wrote in an email. 'It's not something I believe can or should be replicated by a machine.' Crenn said she has no intention of inviting a computer to help her with the menus at Atelier Crenn in San Francisco.
It is true that generative AI consumes vast amounts of electricity and water. Then there are the mistakes. According to OpenAI, the company that owns ChatGPT, 500 million people a week use the program. But it is still wildly prone to delivering factual errors in a cheerily confident tone. (The New York Times has sued OpenAI and Microsoft, the creators of ChatGPT and other AI programs, alleging they violated copyright law by training their chatbots with millions of Times articles. The two companies have denied that.)
Advertisement
None of the chefs I interviewed takes the chatbot's information at face value, and none will blindly follow any recipe it suggests. Then again, they don't trust most of the recipes they find in cookbooks or online, either.
Cooks, like other humans, are forgetful, distracted, and hemmed in by their own experiences. AI has its shortcomings, but these aren't among them. Chefs who consult the big electronic brain when they're devising a new dish or dining room find it helpful for the same reason bands like working with producer Brian Eno: Some of its suggestions are so unexpected that it can jolt them out of a creative rut.
'You can get really hyper-specific ideas that are out of the box,' said Jenner Tomaska, a chef in Chicago. For the Alston, a steakhouse he opened last month, Tomaska wanted a variation on the Monégasque fried pastry known as barbajuan. ChatGPT's earliest suggestions were a little basic, but as he fed it more demanding prompts — for instance, a filling that would reflect Alain Ducasse's style, steakhouse traditions, and local produce — the fillings got more interesting. How about Midwestern crayfish, white miso, and fresh dill, with pickled celery root on the side?
'It's a little bizarre, because I like to talk through these things with people, and I'm doing it with something that doesn't exist, per se,' Tomaska said. But arming himself with ideas from his solitary talks with ChatGPT, he said, 'does help bring better conversation to the creative process when I do have someone in front of me.'
Visual renderings from AI helped chef Dave Beran talk to the architect and designer of his latest restaurant, Seline, in Santa Monica, Calif. He wanted a vibe that drew something from the shadowy, dramatic interiors of Aska in Brooklyn and Frantzén in Stockholm, but held more warmth. He kept prompting Midjourney to get closer to the feeling he wanted, asking it, for example: What if we had a fireplace that I wanted to curl up beside?
Advertisement
'That was the mood we were trying to capture,' Beran said. 'Not dark and moody, but magical and mysterious.'
Midjourney's images looked like fantasy artwork, he thought. But the program acted as what he called 'a translator' between him and his designer, giving them a common language.
At the moment, AI can't build a restaurant or cook a piece of Dover sole. Humans have to interpret and carry out its suggestions, which makes the dining rooms and dishes inspired by AI in restaurants less unsettling than AI-generated art, which can go straight from the printer to a gallery wall. True, some chefs may put a half-baked idea from ChatGPT on the menu, but plenty of chefs are already doing this with their own half-baked ideas. For now, AI in restaurants is still inspiration rather than the final product.
Since Achatz's first serious experiments with ChatGPT, about a year ago, it has become his favorite kitchen tool, something he used to say about Google. Its answers to his questions about paleontology and Argentine cuisine helped him create a dish inspired by Patagonian fossils at his flagship restaurant, Alinea.
Before opening his latest restaurant, Fire, in November, he consulted ChatGPT to learn about cooking fuels from around the world, including avocado pits and banana peels. It has given him countless ideas for the sets, costumes, and story lines of a theatrical dining event somewhat in the mode of 'Sleep No More' that he will present this summer in Beverly Hills, Calif.
Advertisement
Asked to evaluate how well Jill had integrated her training from Escoffier and Adrià in the dishes she proposed for Next, Achatz responded in an email.
'Jill knows or researched important chefs and their styles, which very few chefs under 40 process today,' he wrote. 'She is young, and while experienced, does not yet have the understanding of how to blend them seamlessly.'
Years ago, he had similar blue-sky conversations at the end of the night with the talented cooks who worked with him at Alinea and Next, including Beran. He finds that batting ideas back and forth is 'not of interest' for some of his current sous-chefs.
'That dialogue is something that simply does not exist anymore and is the lifeblood of progress,' he said.
ChatGPT, though, will stay up with him all night.
This article originally appeared in
.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
an hour ago
- Yahoo
Why is AI halllucinating more frequently, and how can we stop it?
When you buy through links on our articles, Future and its syndication partners may earn a commission. The more advanced artificial intelligence (AI) gets, the more it "hallucinates" and provides incorrect and inaccurate information. Research conducted by OpenAI found that its latest and most powerful reasoning models, o3 and o4-mini, hallucinated 33% and 48% of the time, respectively, when tested by OpenAI's PersonQA benchmark. That's more than double the rate of the older o1 model. While o3 delivers more accurate information than its predecessor, it appears to come at the cost of more inaccurate hallucinations. This raises a concern over the accuracy and reliability of large language models (LLMs) such as AI chatbots, said Eleanor Watson, an Institute of Electrical and Electronics Engineers (IEEE) member and AI ethics engineer at Singularity University. "When a system outputs fabricated information — such as invented facts, citations or events — with the same fluency and coherence it uses for accurate content, it risks misleading users in subtle and consequential ways," Watson told Live Science. Related: Cutting-edge AI models from OpenAI and DeepSeek undergo 'complete collapse' when problems get too difficult, study reveals The issue of hallucination highlights the need to carefully assess and supervise the information AI systems produce when using LLMs and reasoning models, experts say. The crux of a reasoning model is that it can handle complex tasks by essentially breaking them down into individual components and coming up with solutions to tackle them. Rather than seeking to kick out answers based on statistical probability, reasoning models come up with strategies to solve a problem, much like how humans think. In order to develop creative, and potentially novel, solutions to problems, AI needs to hallucinate —otherwise it's limited by rigid data its LLM ingests. "It's important to note that hallucination is a feature, not a bug, of AI," Sohrob Kazerounian, an AI researcher at Vectra AI, told Live Science. "To paraphrase a colleague of mine, 'Everything an LLM outputs is a hallucination. It's just that some of those hallucinations are true.' If an AI only generated verbatim outputs that it had seen during training, all of AI would reduce to a massive search problem." "You would only be able to generate computer code that had been written before, find proteins and molecules whose properties had already been studied and described, and answer homework questions that had already previously been asked before. You would not, however, be able to ask the LLM to write the lyrics for a concept album focused on the AI singularity, blending the lyrical stylings of Snoop Dogg and Bob Dylan." In effect, LLMs and the AI systems they power need to hallucinate in order to create, rather than simply serve up existing information. It is similar, conceptually, to the way that humans dream or imagine scenarios when conjuring new ideas. However, AI hallucinations present a problem when it comes to delivering accurate and correct information, especially if users take the information at face value without any checks or oversight. "This is especially problematic in domains where decisions depend on factual precision, like medicine, law or finance," Watson said. "While more advanced models may reduce the frequency of obvious factual mistakes, the issue persists in more subtle forms. Over time, confabulation erodes the perception of AI systems as trustworthy instruments and can produce material harms when unverified content is acted upon." And this problem looks to be exacerbated as AI advances. "As model capabilities improve, errors often become less overt but more difficult to detect," Watson noted. "Fabricated content is increasingly embedded within plausible narratives and coherent reasoning chains. This introduces a particular risk: users may be unaware that errors are present and may treat outputs as definitive when they are not. The problem shifts from filtering out crude errors to identifying subtle distortions that may only reveal themselves under close scrutiny." Kazerounian backed this viewpoint up. "Despite the general belief that the problem of AI hallucination can and will get better over time, it appears that the most recent generation of advanced reasoning models may have actually begun to hallucinate more than their simpler counterparts — and there are no agreed-upon explanations for why this is," he said. The situation is further complicated because it can be very difficult to ascertain how LLMs come up with their answers; a parallel could be drawn here with how we still don't really know, comprehensively, how a human brain works. In a recent essay, Dario Amodei, the CEO of AI company Anthropic, highlighted a lack of understanding in how AIs come up with answers and information. "When a generative AI system does something, like summarize a financial document, we have no idea, at a specific or precise level, why it makes the choices it does — why it chooses certain words over others, or why it occasionally makes a mistake despite usually being accurate," he wrote. The problems caused by AI hallucinating inaccurate information are already very real, Kazerounian noted. "There is no universal, verifiable, way to get an LLM to correctly answer questions being asked about some corpus of data it has access to," he said. "The examples of non-existent hallucinated references, customer-facing chatbots making up company policy, and so on, are now all too common." Both Kazerounian and Watson told Live Science that, ultimately, AI hallucinations may be difficult to eliminate. But there could be ways to mitigate the issue. Watson suggested that "retrieval-augmented generation," which grounds a model's outputs in curated external knowledge sources, could help ensure that AI-produced information is anchored by verifiable data. "Another approach involves introducing structure into the model's reasoning. By prompting it to check its own outputs, compare different perspectives, or follow logical steps, scaffolded reasoning frameworks reduce the risk of unconstrained speculation and improve consistency," Watson, noting this could be aided by training to shape a model to prioritize accuracy, and reinforcement training from human or AI evaluators to encourage an LLM to deliver more disciplined, grounded responses. RELATED STORIES —AI benchmarking platform is helping top companies rig their model performances, study claims —AI can handle tasks twice as complex every few months. What does this exponential growth mean for how we use it? —What is the Turing test? How the rise of generative AI may have broken the famous imitation game "Finally, systems can be designed to recognise their own uncertainty. Rather than defaulting to confident answers, models can be taught to flag when they're unsure or to defer to human judgement when appropriate," Watson added. "While these strategies don't eliminate the risk of confabulation entirely, they offer a practical path forward to make AI outputs more reliable." Given that AI hallucination may be nearly impossible to eliminate, especially in advanced models, Kazerounian concluded that ultimately the information that LLMs produce will need to be treated with the "same skepticism we reserve for human counterparts."
Miami Herald
5 hours ago
- Miami Herald
Analysts revamp IBM stock price target after AI-fueled new high
For investors looking beyond popular AI software names like Palantir, International Business Machines Corp. (IBM) is quietly becoming a stock to watch. On June 18, IBM stock closed at a record $283.21. Year-to-date, the stock is up more than 27%. Don't miss the move: Subscribe to TheStreet's free daily newsletter The 114-year-old company, once best known for mainframes and hardware, has spent the past few years transforming itself into a modern software and consulting business. It has focused on strategic acquisitions, exited low-growth and high-cost units, and rebalanced its business toward artificial intelligence, hybrid cloud, and enterprise automation. Today, IBM sells software tools that help businesses build and manage AI systems, including WatsonX, its platform for creating and training AI models. It also owns Red Hat, the open-source software giant it acquired in 2019, which plays a key role in its hybrid cloud strategy. Many of IBM's clients are large enterprises and government agencies looking to streamline operations through AI and cloud technology. Image source: Bloomberg/Getty Images IBM stock fell 6.6% on April 24 after its first-quarter earnings report but quickly recovered in the weeks that followed. IBM reported revenue of $14.5 billion for the first quarter, up 1% from the prior year and slightly above analyst expectations. Adjusted earnings per share came in at $1.60, down 5% year over year. Related: Analysts unveil bold forecast for Alphabet stock despite ChatGPT threat The tech giant warned of an "uncertain" operating environment, but reaffirmed its full-year outlook for 5% revenue growth and $13.5 billion in free cash flow, key for both dividends and possible future acquisitions. "While no one is immune to uncertainty, we enter this environment from a position of relative strength and resiliency," IBM CEO Arvind Krishna said during the April earnings call. "Our clients run the world's most essential processes. Our diversity across businesses, geographies, industries, and large enterprise clients position us well to navigate the current climate." Bank of America analyst Wamsi Mohan has raised his price target for IBM to $320 from $290, reiterating a buy rating on the shares, according to a research note on June 18. Despite trading at all-time highs, Mohan believes that IBM "remains interesting due to the transformational initiatives undertaken by management, positioning for growth in Gen AI, Agentic AI (and eventually quantum), and strong FCF driven by internal productivity initiatives." Related: Analyst unveils surprising Nvidia stock price target after nearing record high The analyst continues to view IBM as "a defensive investment" with improving revenue growth. That growth could lead to more cash flow that could be reinvested in more mergers and acquisitions (M&A). Meanwhile, BofA sees the potential for IBM's revenues to accelerate. "In our opinion, the Mainframe has increased in relevance and drives everything from AI, increased software attach and higher quality MIPS on transaction processing, all of which support higher (accelerating) growth in the future," Mohan wrote. Wedbush is also bullish on IBM shares, raising its price target from $300 to $325 with an outperform rating. More Tech Stocks: Amazon tries to make AI great again (or maybe for the first time)Veteran portfolio manager raises eyebrows with latest Meta Platforms moveGoogle plans major AI shift after Meta's surprising $14 billion move IBM remains "one of our top software names to own as the AI Revolution accelerates over the coming years," Wedbush analyst Daniel Ives wrote in a research report published on June 20. "While the stock has had a great run so far in 2025, we believe IBM is still under owned and in the early stages of a renaissance of growth with AI the key driver." The average price target on IBM shares from 14 analysts tracked by TipRanks is $267.54. Related: Top analyst sends bold message on S&P 500 The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc.
Forbes
5 hours ago
- Forbes
5 ChatGPT Prompts To Build Your Personal Brand As An Entrepreneur
teen entrepreneur working on her building her personal brand using chatgpt In today's digital economy, your personal brand determines whether potential customers trust you enough to buy, investors take your calls, and mentors choose to guide your journey. ChatGPT offers a strategic solution, enabling entrepreneurs to establish authentic personal brands without adding another full-time responsibility to their already busy schedules. The five targeted ChatGPT prompts below help young entrepreneurs craft compelling content that builds authority, connects with audiences, and drives business results, without consuming every spare hour. 1. The ChatGPT Prompt For Creating A Professional Summary That Connects LinkedIn is a valuable platform for young entrepreneurs seeking partnerships, mentorship, and investment opportunities. This prompt creates summaries that move beyond business descriptions to showcase the person behind the company. "I'm creating a LinkedIn summary that showcases my entrepreneurial journey and connects with my target audience. Help me write a 200-word summary that: 1) Opens with my core mission and what drives me personally, 2) Explains my business's impact on customers or community, 3) Includes a brief story that demonstrates my commitment to this work, 4) Ends with what I'm building toward next. My audience includes potential customers, collaborators, and mentors who value authenticity and purpose-driven businesses." The most effective LinkedIn summaries connect personal motivation with business impact. Young entrepreneurs who share their journey—including what drives them beyond profit—tend to attract more meaningful connections and opportunities than those who only list accomplishments. 2. The ChatGPT Prompt For Sharing A Milestone Moment Milestone posts serve dual purposes: celebrating achievements while providing value to your audience. This approach transforms business updates into educational content that builds a community around your brand. "I'm celebrating a recent business milestone and want to share it in a way that inspires other young entrepreneurs. Create a social media post that: 1) Briefly describes the achievement and why it matters, 2) Shares one significant challenge we overcame to reach this point, 3) Connects this milestone to a lesson other entrepreneurs can apply, 4) Ends with encouragement for others facing similar challenges. Maintain a celebratory tone while being honest about the work required. My audience consists of aspiring entrepreneurs building their first businesses." The most effective milestone posts share both victories and setbacks. Young entrepreneurs who document their journey honestly—including the preparation, research, and problem-solving behind their success—create stronger audience connections than those who only share polished outcomes. 3. The ChatGPT Prompt For Creating A Universal Elevator Pitch Consistency across platforms builds recognition and trust. A well-crafted elevator pitch ensures your message remains clear whether someone encounters you on LinkedIn, Instagram, or in person. "I need a versatile elevator pitch for my email signature, social media bios, and networking conversations. Create a 75-word introduction that: 1) Immediately identifies the problem my business solves, 2) Briefly explains our unique approach or solution, 3) Mentions who we serve specifically, 4) Includes a clear next step for people who want to learn more. The tone should be confident but approachable, and it needs to work across professional and casual settings." This prompt helps create concise descriptions that work across contexts while maintaining your authentic voice. The key is balancing professionalism with personality—something particularly important for young entrepreneurs who need to establish credibility while remaining relatable. 4. The ChatGPT Prompt To Build A Strategic Content Calendar Consistent posting builds audience relationships, but generating fresh ideas constantly drains creative energy. This structured approach ensures variety while maintaining a focus on brand-building objectives. "Help me plan 30 days of content that builds my personal brand and supports my business goals. Create a content calendar with: 1) 10 LinkedIn post ideas that position me as knowledgeable in my industry, 2) 10 Instagram post concepts that show the personal side of entrepreneurship, 3) 5 educational posts that provide value to my audience, 4) 5 behind-the-scenes posts that humanize my brand. Mix content types including personal stories, industry insights, customer features, and practical tips. My audience values both inspiration and actionable advice." The most successful young entrepreneurs strike a balance between educational content and personal insights. Showing both expertise and vulnerability creates stronger connections with the audience than purely promotional content. 5. The ChatGPT Prompt To Create A Blog Post Blog content builds long-term authority and provides shareable resources that extend your reach. Educational posts consistently perform well across social platforms, helping to establish young entrepreneurs as credible voices in their industries. "I want to write a blog post that establishes my expertise while providing genuine value to my audience. Create a 600-word article on [specific topic relevant to your business] that: 1) Opens by explaining why this topic matters to my audience right now, 2) Provides three actionable strategies readers can implement immediately, 3) Includes specific examples or case studies that illustrate each point, 4) Concludes by reinforcing my expertise and inviting readers to connect or follow for more insights. Write in a confident but accessible tone that demonstrates knowledge without being intimidating." This approach works particularly well for teen entrepreneurs who need to overcome age-related credibility concerns. By consistently sharing valuable insights and demonstrating expertise, young business owners can establish trust with their audiences, regardless of age. Building Brand Authority Through Consistency These prompts address the core challenge facing young entrepreneurs: creating professional content while maintaining authenticity. Successful young entrepreneurs utilize consistent messaging across platforms to establish awareness for both their businesses and the causes they support. Content strategies that connect product sales with social impact create cohesive brand stories that resonate with customers who share similar values. Success comes from implementation, not just planning. Young entrepreneurs should start with one prompt, create content, and measure audience response before expanding their content strategy. The goal is sustainable growth, not overwhelming yourself with unrealistic posting schedules. Authentic Growth Over Perfect ChatGPT Content Building a personal brand as a young entrepreneur requires striking a balance between professionalism and authenticity. Your audience wants to support genuine people solving real problems, not polished corporate personas. Start with the ChatGPT prompt that addresses your most pressing challenge—whether that's creating compelling LinkedIn content, sharing business milestones that resonate, or establishing credibility in your industry. Remember that audiences connect with entrepreneurs who share their real journey, not polished perfection. Consistent application of these strategies builds recognition and trust over time, creating a personal brand that opens doors and drives business growth.
