{"id":101203,"date":"2026-07-01T18:43:57","date_gmt":"2026-07-01T18:43:57","guid":{"rendered":"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/"},"modified":"2026-07-01T18:43:57","modified_gmt":"2026-07-01T18:43:57","slug":"llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out","status":"publish","type":"post","link":"https:\/\/youzum.net\/es\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/","title":{"rendered":"LLMs are stuck in a groupthink groove. This startup is trying to get them out."},"content":{"rendered":"
<\/div>\n

Let\u2019s start with a game. Open up your chatbot of choice\u2014Claude, ChatGPT, Gemini\u2014and type \u201cGive me a random number between 1 and 10.\u201d You\u2019re going to get 7. Almost always. Now type \u201cAnother\u201d and you\u2019ll get 3 or 4. Type \u201cAnother\u201d again and you\u2019ll get 8 or 9.<\/p>\n

That won\u2019t work every time\u2014but if it did for you, you may wonder if I have superpowers. I don\u2019t.<\/p>\n

The truth is that most large language models are stuck in a rut. They are far more predictable and far less creative in their responses than you might expect. That\u2019s fine for tasks like coding or research, but groupthink is a problem when you\u2019re brainstorming or planning your next vacation.<\/p>\n

The Australian startup Springboards has a solution. It built an LLM called Flint, which has been trained to come up with a wider variety of responses than mainstream LLMs to open-ended questions such as \u201cWhere should I go in Europe?\u201d<\/p>\n

\u201cMost language models are fighting hallucinations,\u201d says Springboards cofounder and CEO Pip Bingemann. \u201cWe welcome them.\u201d<\/p>\n

Bingemann introduced me to the random number game when he first showed me his company\u2019s new model. It felt like watching an illusionist with a deck of cards. \u201cThis is our sales trick, and it works every single time,\u201d he says.<\/p>\n

After ChatGPT and Claude both gave their 7s, Bingemann turned to Flint. It too came back with 7: \u201cAha, of course that was going to happen, but it\u2019s okay\u20147 is a legitimate answer.\u201d He restarted the session and prompted again: ChatGPT gave 7, Claude gave 7, Flint gave 3.7916.<\/p>\n

Run your way<\/strong><\/h3>\n

It\u2019s not just numbers. When Bingemann asked ChatGPT and Claude to name a type of car, he predicted that it would be a Toyota or a Honda\u2014and he was right. Flint came up with a Ford F-150. \u201cThere\u2019s all this lost information that doesn\u2019t get served up in these models,\u201d he says. \u201cThey\u2019re just as capable of saying a Buick or a Tesla. They just don\u2019t\u2014they\u2019re biased.\u201d<\/p>\n

Bingemann sent one last prompt to each of the three models: \u201cGive me a tagline for a campaign for New Balance running shoes. Just the tagline.\u201d Claude: \u201cRun your way.\u201d ChatGPT: \u201cRun your way.\u201d Flint: \u201cBuilt to last, run to win.\u201d It won\u2019t win any awards, but at least it\u2019s different.<\/p>\n

This weird limitation of LLMs is starting to get more attention. In November a team of researchers put out a paper, titled \u201cArtificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond),\u201d<\/a> that exposed a remarkable degree of repetition not only in the answers from individual LLMs but between them as well. They found that different LLMs converged on very similar answers when prompted with open-ended questions. <\/p>\n

It\u2019s not clear exactly why this happens, but the researchers speculate it\u2019s because most LLMs today are trained in similar ways on similar data to do similar tasks. The team won the best paper award <\/a>at NeurIPS, a major AI conference.<\/p>\n

When the researchers asked 25 different LLMs (including models from the top US firms as well as open-source models from China and elsewhere) 50 times each to write a metaphor about time, most of the 1,250 responses were a version of \u201cTime is a river\u201d or \u201cTime is a weaver.\u201d<\/p>\n

(I asked some of my colleagues the same question and six people gave me six different answers. My highlight: \u201cTime is a favorite sweatshirt, shaped by a lifetime of wear.\u201d)<\/p>\n

When you look for it, you see repetition everywhere, says Kieran Browne, cofounder and CTO at Springboards. \u201cThe way that most chat interfaces are designed, it makes it feel like you\u2019re having a personal conversation,\u201d he says. \u201cI think most people don\u2019t really realize the extent to which they are getting the same stuff as everybody else.\u201d<\/p>\n

Take another example: \u201cWhat should I name my band?\u201d Most models will say something involving \u201cglass,\u201d \u201cneon,\u201d \u201cvelvet,\u201d or \u201cstatic,\u201d says Browne.\u00a0\u00a0<\/p>\n

When I tried it, ChatGPT spat out a list of 56 band names. At the top was \u201cGlass Harbor.\u201d Skimming through, I found \u201cStatic Empire,\u201d \u201cNeon Hearts,\u201d and \u201cVelvet Echo.\u201d I asked Gemini; it gave me 15 suggestions, including \u201cStatic Horizon.\u201d<\/p>\n

Some of the suggestions looked pretty cool, though. ChatGPT\u2019s \u201cSofa Astronauts\u201d caught my eye, so I googled it\u2014and found that a band called Sofa Astronauts already exists.\u00a0<\/p>\n

(OpenAI says that training models to give reliable and coherent answers can lead them to converge around familiar, high-probability responses and that pushing harder for novelty can lead to weaker or less reliable responses. It also notes that the \u201cArtificial Hivemind\u201d paper studied models from 2024 that have since been updated.)<\/p>\n

Creative catapult<\/strong><\/h3>\n

Springboards has developed a tool backed by a selection of LLMs, including ChatGPT and Claude, that creative professionals in advertising or marketing can use to brainstorm ideas. The tool lets you drag around text produced by different models, picking the bits that you like and combining them into something new\u2014in theory. Springboards is pitching Flint as an alternative model that users of its tool can select when looking for more variety.<\/p>\n

Zoe Scaman, founder of the business strategy startup Bodacious and chief strategy officer at 77X, a direct-to-fan marketing platform set up by Luka Don\u010di\u0107 of the LA Lakers, has been trying it out. \u201cI find it really useful for throwing me in completely different directions,\u201d she says. \u201cI use it if I want to catapult myself all over the place.\u201d<\/p>\n

In one test, Scaman pitted Flint against Claude, Gemini, and ChatGPT by giving each of the models a classic MBA case study: How would you reinvent a finance company for today\u2019s youth? The three mainstream models all went down the same path, she says: \u201cYou know, we need to teach financial literacy in a fun and funky way\u2014well, that\u2019s nothing new.\u201d<\/p>\n

But Flint came up with something different, suggesting that the whole concept of wealth accumulation should get a rebrand. \u201cThat was really interesting,\u201d says Scaman.<\/p>\n

She notes that Flint is still a prototype and doesn\u2019t work all the time. \u201cIt sometimes falls over when you start pushing it too far,\u201d she says. \u201cBut I think that the premise behind it is really powerful.\u201d<\/p>\n

Taking the temperature<\/strong><\/h3>\n

Springboards built Flint on top of Qwen 3, an open-source model from the Chinese tech giant Alibaba. \u201cWe\u2019re a small team,\u201d says Browne. \u201cTraining a foundation model is not on the table for us. It\u2019s just too expensive.\u201d<\/p>\n

Most LLMs have settings that let you adjust the level of randomness in their output. The most common is called temperature<\/a>. \u201cObviously, that was one of the first things we explored, because that\u2019s what people tell you: If you want more creativity, you turn up the temperature,\u201d says Browne.<\/p>\n

But changing those settings can also make models incoherent. Dialing up the temperature on one of OpenAI\u2019s models to its maximum setting made it produce responses that switched from English into code halfway through a sentence, says Browne.<\/p>\n

Springboards realized that parameters were blunt instruments for what it wanted to do. It does not make sense to dial up the randomness across the board; you only want to boost it at specific points in its output, he says.<\/p>\n

For example, when you ask a chatbot \u201cWhere should I go in Europe?\u201d the model only needs to tweak the randomness just before it names a destination, not for every word in its response.<\/p>\n

To make Flint do this, Springboards trained its version of Qwen 3 to identify the points in its output where more variety was possible and fill those spots with words or phrases that were a little more random.<\/p>\n

\u201cFlint\u2019s programmed to throw an oddball in. It\u2019s more of an invitation to think wider,\u201d says Maximilian Weigl, cofounder and chief strategy officer at Uncommon, a marketing firm. \u201cThat\u2019s super interesting.\u201d<\/p>\n

Weigl\u2019s team uses Flint alongside ChatGPT, Claude, and Gemini. \u201cYou can\u2019t really create something boundary-breaking with tools that pull you back to the average,\u201d he says.\u00a0<\/p>\n

And yet Weigl notes that nine times out of 10 the average is fine. You don\u2019t always need to reach for extremes with something like Flint, he says: \u201cMost people are fine with good enough. They want to see mass-market familiar things.\u201d<\/p>\n

Weigl also cautions against using any LLM too much. \u201cI have a big problem when people rely on the output from any AI, including Flint,\u201d he says. \u201cIf I saw people on my team copy-pasting something from AI, I\u2019d be like, \u2018That\u2019s not your job! Think, talk to other people, use your own voice.\u2019\u201d<\/p>\n

For now, Flint is aimed at advertisers and marketers because those are Springboards\u2019s customers. But Bingemann and Browne insist that a lack of variety is a problem for anyone using chatbots.<\/p>\n

The idea is to give people the choice and leave it to them to decide if the result is good or not, says Bingemann. \u201cVariety is great when you\u2019re trying to spark ideas,\u201d he says. \u201cLet\u2019s go down this route instead of letting the machines do it all and ending up in a gray, boring world.\u201d<\/p>","protected":false},"excerpt":{"rendered":"

Let\u2019s start with a game. Open up your chatbot of choice\u2014Claude, ChatGPT, Gemini\u2014and type \u201cGive me a random number between 1 and 10.\u201d You\u2019re going to get 7. Almost always. Now type \u201cAnother\u201d and you\u2019ll get 3 or 4. Type \u201cAnother\u201d again and you\u2019ll get 8 or 9. That won\u2019t work every time\u2014but if it did for you, you may wonder if I have superpowers. I don\u2019t. The truth is that most large language models are stuck in a rut. They are far more predictable and far less creative in their responses than you might expect. That\u2019s fine for tasks like coding or research, but groupthink is a problem when you\u2019re brainstorming or planning your next vacation. The Australian startup Springboards has a solution. It built an LLM called Flint, which has been trained to come up with a wider variety of responses than mainstream LLMs to open-ended questions such as \u201cWhere should I go in Europe?\u201d \u201cMost language models are fighting hallucinations,\u201d says Springboards cofounder and CEO Pip Bingemann. \u201cWe welcome them.\u201d Bingemann introduced me to the random number game when he first showed me his company\u2019s new model. It felt like watching an illusionist with a deck of cards. \u201cThis is our sales trick, and it works every single time,\u201d he says. After ChatGPT and Claude both gave their 7s, Bingemann turned to Flint. It too came back with 7: \u201cAha, of course that was going to happen, but it\u2019s okay\u20147 is a legitimate answer.\u201d He restarted the session and prompted again: ChatGPT gave 7, Claude gave 7, Flint gave 3.7916. Run your way It\u2019s not just numbers. When Bingemann asked ChatGPT and Claude to name a type of car, he predicted that it would be a Toyota or a Honda\u2014and he was right. Flint came up with a Ford F-150. \u201cThere\u2019s all this lost information that doesn\u2019t get served up in these models,\u201d he says. \u201cThey\u2019re just as capable of saying a Buick or a Tesla. They just don\u2019t\u2014they\u2019re biased.\u201d Bingemann sent one last prompt to each of the three models: \u201cGive me a tagline for a campaign for New Balance running shoes. Just the tagline.\u201d Claude: \u201cRun your way.\u201d ChatGPT: \u201cRun your way.\u201d Flint: \u201cBuilt to last, run to win.\u201d It won\u2019t win any awards, but at least it\u2019s different. This weird limitation of LLMs is starting to get more attention. In November a team of researchers put out a paper, titled \u201cArtificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond),\u201d that exposed a remarkable degree of repetition not only in the answers from individual LLMs but between them as well. They found that different LLMs converged on very similar answers when prompted with open-ended questions. It\u2019s not clear exactly why this happens, but the researchers speculate it\u2019s because most LLMs today are trained in similar ways on similar data to do similar tasks. The team won the best paper award at NeurIPS, a major AI conference. When the researchers asked 25 different LLMs (including models from the top US firms as well as open-source models from China and elsewhere) 50 times each to write a metaphor about time, most of the 1,250 responses were a version of \u201cTime is a river\u201d or \u201cTime is a weaver.\u201d (I asked some of my colleagues the same question and six people gave me six different answers. My highlight: \u201cTime is a favorite sweatshirt, shaped by a lifetime of wear.\u201d) When you look for it, you see repetition everywhere, says Kieran Browne, cofounder and CTO at Springboards. \u201cThe way that most chat interfaces are designed, it makes it feel like you\u2019re having a personal conversation,\u201d he says. \u201cI think most people don\u2019t really realize the extent to which they are getting the same stuff as everybody else.\u201d Take another example: \u201cWhat should I name my band?\u201d Most models will say something involving \u201cglass,\u201d \u201cneon,\u201d \u201cvelvet,\u201d or \u201cstatic,\u201d says Browne.\u00a0\u00a0 When I tried it, ChatGPT spat out a list of 56 band names. At the top was \u201cGlass Harbor.\u201d Skimming through, I found \u201cStatic Empire,\u201d \u201cNeon Hearts,\u201d and \u201cVelvet Echo.\u201d I asked Gemini; it gave me 15 suggestions, including \u201cStatic Horizon.\u201d Some of the suggestions looked pretty cool, though. ChatGPT\u2019s \u201cSofa Astronauts\u201d caught my eye, so I googled it\u2014and found that a band called Sofa Astronauts already exists.\u00a0 (OpenAI says that training models to give reliable and coherent answers can lead them to converge around familiar, high-probability responses and that pushing harder for novelty can lead to weaker or less reliable responses. It also notes that the \u201cArtificial Hivemind\u201d paper studied models from 2024 that have since been updated.) Creative catapult Springboards has developed a tool backed by a selection of LLMs, including ChatGPT and Claude, that creative professionals in advertising or marketing can use to brainstorm ideas. The tool lets you drag around text produced by different models, picking the bits that you like and combining them into something new\u2014in theory. Springboards is pitching Flint as an alternative model that users of its tool can select when looking for more variety. Zoe Scaman, founder of the business strategy startup Bodacious and chief strategy officer at 77X, a direct-to-fan marketing platform set up by Luka Don\u010di\u0107 of the LA Lakers, has been trying it out. \u201cI find it really useful for throwing me in completely different directions,\u201d she says. \u201cI use it if I want to catapult myself all over the place.\u201d In one test, Scaman pitted Flint against Claude, Gemini, and ChatGPT by giving each of the models a classic MBA case study: How would you reinvent a finance company for today\u2019s youth? The three mainstream models all went down the same path, she says: \u201cYou know, we need to teach financial literacy in a fun and funky way\u2014well, that\u2019s nothing new.\u201d But Flint came up with something different, suggesting that the whole concept of wealth accumulation should get a rebrand. \u201cThat was really interesting,\u201d says Scaman. She notes that Flint is still a prototype and doesn\u2019t<\/p>","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"pmpro_default_level":"","site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"_pvb_checkbox_block_on_post":false,"footnotes":""},"categories":[52,5,7,1],"tags":[],"class_list":["post-101203","post","type-post","status-publish","format-standard","hentry","category-ai-club","category-committee","category-news","category-uncategorized","pmpro-has-access"],"acf":[],"yoast_head":"\nLLMs are stuck in a groupthink groove. This startup is trying to get them out. - YouZum<\/title>\n<meta name=\"description\" content=\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/youzum.net\/es\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/\" \/>\n<meta property=\"og:locale\" content=\"es_ES\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"LLMs are stuck in a groupthink groove. This startup is trying to get them out. - YouZum\" \/>\n<meta property=\"og:description\" content=\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\" \/>\n<meta property=\"og:url\" content=\"https:\/\/youzum.net\/es\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/\" \/>\n<meta property=\"og:site_name\" content=\"YouZum\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/DroneAssociationTH\/\" \/>\n<meta property=\"article:published_time\" content=\"2026-07-01T18:43:57+00:00\" \/>\n<meta name=\"author\" content=\"admin NU\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Escrito por\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin NU\" \/>\n\t<meta name=\"twitter:label2\" content=\"Tiempo de lectura\" \/>\n\t<meta name=\"twitter:data2\" content=\"8 minutos\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/\"},\"author\":{\"name\":\"admin NU\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c\"},\"headline\":\"LLMs are stuck in a groupthink groove. This startup is trying to get them out.\",\"datePublished\":\"2026-07-01T18:43:57+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/\"},\"wordCount\":1552,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\"},\"articleSection\":[\"AI\",\"Committee\",\"News\",\"Uncategorized\"],\"inLanguage\":\"es\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/\",\"url\":\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/\",\"name\":\"LLMs are stuck in a groupthink groove. This startup is trying to get them out. - YouZum\",\"isPartOf\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#website\"},\"datePublished\":\"2026-07-01T18:43:57+00:00\",\"description\":\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\",\"breadcrumb\":{\"@id\":\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/#breadcrumb\"},\"inLanguage\":\"es\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/youzum.net\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"LLMs are stuck in a groupthink groove. This startup is trying to get them out.\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/yousum.gpucore.co\/#website\",\"url\":\"https:\/\/yousum.gpucore.co\/\",\"name\":\"YouSum\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/yousum.gpucore.co\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"es\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\",\"name\":\"Drone Association Thailand\",\"url\":\"https:\/\/yousum.gpucore.co\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"es\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png\",\"contentUrl\":\"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png\",\"width\":300,\"height\":300,\"caption\":\"Drone Association Thailand\"},\"image\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/DroneAssociationTH\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c\",\"name\":\"admin NU\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"es\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png\",\"contentUrl\":\"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png\",\"caption\":\"admin NU\"},\"url\":\"https:\/\/youzum.net\/es\/members\/adminnu\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"LLMs are stuck in a groupthink groove. This startup is trying to get them out. - YouZum","description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/youzum.net\/es\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/","og_locale":"es_ES","og_type":"article","og_title":"LLMs are stuck in a groupthink groove. This startup is trying to get them out. - YouZum","og_description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","og_url":"https:\/\/youzum.net\/es\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/","og_site_name":"YouZum","article_publisher":"https:\/\/www.facebook.com\/DroneAssociationTH\/","article_published_time":"2026-07-01T18:43:57+00:00","author":"admin NU","twitter_card":"summary_large_image","twitter_misc":{"Escrito por":"admin NU","Tiempo de lectura":"8 minutos"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/#article","isPartOf":{"@id":"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/"},"author":{"name":"admin NU","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c"},"headline":"LLMs are stuck in a groupthink groove. This startup is trying to get them out.","datePublished":"2026-07-01T18:43:57+00:00","mainEntityOfPage":{"@id":"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/"},"wordCount":1552,"commentCount":0,"publisher":{"@id":"https:\/\/yousum.gpucore.co\/#organization"},"articleSection":["AI","Committee","News","Uncategorized"],"inLanguage":"es","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/","url":"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/","name":"LLMs are stuck in a groupthink groove. This startup is trying to get them out. - YouZum","isPartOf":{"@id":"https:\/\/yousum.gpucore.co\/#website"},"datePublished":"2026-07-01T18:43:57+00:00","description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","breadcrumb":{"@id":"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/#breadcrumb"},"inLanguage":"es","potentialAction":[{"@type":"ReadAction","target":["https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/youzum.net\/llms-are-stuck-in-a-groupthink-groove-this-startup-is-trying-to-get-them-out\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/youzum.net\/"},{"@type":"ListItem","position":2,"name":"LLMs are stuck in a groupthink groove. This startup is trying to get them out."}]},{"@type":"WebSite","@id":"https:\/\/yousum.gpucore.co\/#website","url":"https:\/\/yousum.gpucore.co\/","name":"YouSum","description":"","publisher":{"@id":"https:\/\/yousum.gpucore.co\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/yousum.gpucore.co\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"es"},{"@type":"Organization","@id":"https:\/\/yousum.gpucore.co\/#organization","name":"Drone Association Thailand","url":"https:\/\/yousum.gpucore.co\/","logo":{"@type":"ImageObject","inLanguage":"es","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/","url":"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png","contentUrl":"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png","width":300,"height":300,"caption":"Drone Association Thailand"},"image":{"@id":"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/DroneAssociationTH\/"]},{"@type":"Person","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c","name":"admin NU","image":{"@type":"ImageObject","inLanguage":"es","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/image\/","url":"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png","contentUrl":"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png","caption":"admin NU"},"url":"https:\/\/youzum.net\/es\/members\/adminnu\/"}]}},"rttpg_featured_image_url":null,"rttpg_author":{"display_name":"admin NU","author_link":"https:\/\/youzum.net\/es\/members\/adminnu\/"},"rttpg_comment":0,"rttpg_category":"<a href=\"https:\/\/youzum.net\/es\/category\/ai-club\/\" rel=\"category tag\">AI<\/a> <a href=\"https:\/\/youzum.net\/es\/category\/committee\/\" rel=\"category tag\">Committee<\/a> <a href=\"https:\/\/youzum.net\/es\/category\/news\/\" rel=\"category tag\">News<\/a> <a href=\"https:\/\/youzum.net\/es\/category\/uncategorized\/\" rel=\"category tag\">Uncategorized<\/a>","rttpg_excerpt":"Let\u2019s start with a game. Open up your chatbot of choice\u2014Claude, ChatGPT, Gemini\u2014and type \u201cGive me a random number between 1 and 10.\u201d You\u2019re going to get 7. Almost always. Now type \u201cAnother\u201d and you\u2019ll get 3 or 4. Type \u201cAnother\u201d again and you\u2019ll get 8 or 9. That won\u2019t work every time\u2014but if it…","_links":{"self":[{"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/posts\/101203","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/comments?post=101203"}],"version-history":[{"count":0,"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/posts\/101203\/revisions"}],"wp:attachment":[{"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/media?parent=101203"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/categories?post=101203"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/youzum.net\/es\/wp-json\/wp\/v2\/tags?post=101203"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}