<\/p>\n
It turns out this process works on text too. Inception has trained its LLMs to take a random string of words and turn it into sentences that make sense. Diffusion LLMs still use transformers to encode meaning, but by producing whole blocks of text at once, they make transformers do more for less. \u201cYou\u2019re still using a big transformer model, but you can predict many tokens at the same time,\u201d says Inception\u2019s cofounder and CEO, Stefano Ermon. \u201cThat\u2019s why these models are so much faster and cost-efficient compared to what most other people are building today.\u201d<\/p>\n
The challenge was to take a technology designed for image generation and apply it to text. With images, if you need to change a blue pixel to a red one you can step through intermediate colors, says Ermon. That doesn\u2019t work with text: \u201cWhen you have \u2018cat\u2019 and \u2018dog,\u2019 there is not really something in between.\u201d\u00a0<\/p>\n
Ermon is also a researcher at Stanford University. In 2024, he and a pair of his Stanford colleagues figured out the math to make diffusion models work with text. They trained a diffusion model that matched the performance of GPT-2\u2014an LLM that OpenAI built in 2019\u2014but was 10 times faster. It was enough for Ermon to spin out a company.\u00a0<\/p>\n
Today he has his sights on the big league. Inception claims its latest model, Mercury 2, performs as well as some of OpenAI\u2019s GPT-4 models, released in 2023, but again 10 times faster. \u201cWe\u2019re bullish about this approach because it\u2019s the one that is going to scale up,\u201d says Ermon.\u00a0<\/p>\n
The only things that matter are speed and cost, he adds: \u201cUltimately, the currency is going to be intelligence per dollar.\u201d<\/p>\n
Inception is not the only company betting on diffusion. Google is also experimenting with this approach and has built a prototype LLM called Diffusion Gemma. But Ermon is not worried about the competition. \u201cI think it’s validating,\u201d he says. \u201cThis is the future.\u201d<\/p>\n
Pathway, another startup based in Palo Alto, is perhaps the most extreme of this new bunch. It wants to free LLMs from the constraints of language.<\/p>\n
The firm has built a type of LLM called Dragon Hatchling (named after the dragons in Terry Pratchett\u2019s novel Color of Magic<\/em>, which materialize if you think about them hard enough). Its standout result so far is a high score on a benchmark that pits LLMs against more than 250,000 very hard sudoku puzzles. Dragon Hatchling beat more than 97% of the puzzles; several leading LLMs from the top labs failed to solve any.\u00a0<\/p>\n<\/p><\/div>\n #startups #chasing #big #LLMs<\/p>\n","protected":false},"excerpt":{"rendered":" It turns out this process works on text too. Inception has trained its LLMs to…<\/p>\n","protected":false},"author":1,"featured_media":19630,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[7],"tags":[152,1680,7507,7503],"class_list":["post-19629","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-stories","tag-big","tag-chasing","tag-llms","tag-startups"],"featured_image_urls":{"full":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs.jpg",1200,600,false],"thumbnail":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs-150x150.jpg",150,150,true],"medium":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs-300x150.jpg",300,150,true],"medium_large":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs-768x384.jpg",640,320,true],"large":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs-1024x512.jpg",640,320,true],"1536x1536":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs.jpg",1200,600,false],"2048x2048":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs.jpg",1200,600,false],"covernews-featured":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs-1024x512.jpg",1024,512,true],"covernews-medium":["https:\/\/news678.top\/wp-content\/uploads\/2026\/08\/269798_futureLLMs-540x340.jpg",540,340,true]},"author_info":{"display_name":"admin","author_link":"https:\/\/news678.top\/?author=1"},"category_info":"Stories<\/a>","tag_info":"Stories","comment_count":"0","_links":{"self":[{"href":"https:\/\/news678.top\/index.php?rest_route=\/wp\/v2\/posts\/19629","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/news678.top\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/news678.top\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/news678.top\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/news678.top\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=19629"}],"version-history":[{"count":0,"href":"https:\/\/news678.top\/index.php?rest_route=\/wp\/v2\/posts\/19629\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/news678.top\/index.php?rest_route=\/wp\/v2\/media\/19630"}],"wp:attachment":[{"href":"https:\/\/news678.top\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=19629"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/news678.top\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=19629"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/news678.top\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=19629"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}