Richard Socher at the World Economic Forum\u2019s Annual Meeting of the New Champions in Dalian, China, in June 2024 Credit: World Economic Forum\/Jakob Polacsek \/ CC BY-NC-SA 2.0 via Flickr, cropped Recursive Superintelligence, the startup trying to build AI that improves itself, has released its first model. Flint is a free coding model small enough to run on a laptop. Recursive says its own automated research system helped build nearly every part of it. The company announced Flint on Wednesday and published the weights on Hugging Face under the Apache 2.0 licence, which allows free use and changes.<\/p>\n
Recursive calls the system its Eureka Machine. It proposes ideas, builds them, tests them and uses the results to choose what to try next. A small team of people and agents built Flint from scratch in under four months, the company said. How the AI helped build it The Eureka Machine worked on the data, the training recipes, the infrastructure and the code that runs the model, according to Recursive.<\/p>\n
It also searched for the model\u2019s shape. Flint has 40 billion parameters, but only 2.5 billion of them work on each word or code fragment. That design, called mixture of experts, keeps it light enough for a laptop. The system then compressed the model\u2019s output layer to 4 bits, which Recursive says goes \u201cagainst common practice\u201d.<\/p>\n
It also wrote custom code to run Flint faster on Apple and Nvidia chips. Chief executive Richard Socher said on X that the low-bit work shrank Flint by 71% with almost no loss in quality, TechCrunch reported. Nvidia said in a blog post that this code sped up Flint on its RTX Spark machines by 55%, without changing the model itself. Fast on a MacBook Recursive says Flint produces more than 125 tokens, or chunks of text, per second on a MacBook Pro with an M4 Pro chip.<\/p>\n
On Nvidia\u2019s RTX Spark laptops it tops 150. In the company\u2019s tests, similar models such as Google\u2019s Gemma 4 managed 58 to 69 on the same Mac, TechCrunch reported. In Recursive\u2019s own tests, Flint scored 74.2% on SWE-bench Verified, which measures how well AI fixes real software bugs. It scored 54.9% on Terminal-Bench 2.1.<\/p>\n
The company says it leads mixture-of-experts models under 50 billion parameters on these coding tests. TNW has not verified the scores, and Recursive\u2019s model card notes that test setups vary. Recursive also says Flint may need more guidance than larger closed coding models. It suggests breaking big tasks into smaller steps.<\/p>\n
The model card warns that Flint can still produce wrong or harmful content and faulty code. \u201cIt suggests that fully recursive self-improvement, or RSI \u2026 is getting closer,\u201d Will Knight wrote for Wired after testing Flint. The self-improvement bet Recursive came out of stealth in May after it raised $650 million. TechCrunch puts its valuation at $4.65bn, with GV, Nvidia and AMD Ventures among its backers. It later signed a $410M AWS deal for computing power. \u201cOur goal at Recursive is to build recursively self-improving superintelligence to automate knowledge discovery,\u201d the company wrote.<\/p>\n
Socher told Wired the long-term aim is full recursive self-improvement, with humans overseeing the process and able to stop it, according to TechCrunch. Larger labs are moving the same way. Sergey Brin said last week that Gemini now helps build the next Gemini. The idea is contested.<\/p>\n
Some researchers fear AI that improves itself could slip out of human control. In the US, Ro Khanna has planned a bill to ban self-improving AI. Get the TNW newsletter \u2715 Close Get the TNW newsletter Get the most important tech news in your inbox each week.<\/p>\n","protected":false},"excerpt":{"rendered":"
Richard Socher at the World Economic Forum\u2019s Annual Meeting of the New Champions in Dalian, China, in June 2024 Credit: World Economic Forum\/Jakob Polacsek \/ CC BY-NC-SA 2.0 via Flickr, cropped Recursive Superintelligence, the startup\u2026<\/p>\n","protected":false},"author":1,"featured_media":30887,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[],"class_list":["post-30886","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technology"],"_links":{"self":[{"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/posts\/30886","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/comments?post=30886"}],"version-history":[{"count":0,"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/posts\/30886\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/media\/30887"}],"wp:attachment":[{"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/media?parent=30886"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/categories?post=30886"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/demo.tcjhomeaccents.com\/wp\/wp-json\/wp\/v2\/tags?post=30886"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}