heise-offer: Build and run AI projects: Submit now to Mastering AI
The article announces the upcoming 'Mastering AI' conference scheduled to take place in Karlsruhe on April 28th and 29th, 2027. It mentions that the call for proposals for the specialized conference on AI projects is open until November 19th, 2023. The event appears to focus on practical applications and development of artificial intelligence projects.
Aleph Alpha, a German startup based in Heidelberg, has released a new large language model called "Kolibri." Designed specifically for business-critical applications in public administration and industry, the model was developed with a focus on the German language. Twenty-three percent of its pre-training data consisted of German texts, and its tokenizer, software that breaks text into smaller units, is optimized for German. This optimization could make processing German content more efficient and cost-effective. The model contains 78 billion parameters, making it relatively small compared to other models. Parameters are numerical values learned during training that determine how inputs are converted into outputs. According to Ilhan Scheer, co-founder of Aleph Alpha, the company aims to provide the best model for German-speaking customers in the public sector, legal field, and industry, not to compete with larger all-purpose models. Aleph Alpha will soon cease operating as an independent entity. It is being acquired by Cohere, a much larger Canadian competitor. After the acquisition, the company will operate under the name Cohere globally. Aleph Alpha's location in Heidelberg will remain and become a "tech hub," while Berlin will join Toronto as a second corporate center for the new Canada-German enterprise. Regulatory approvals are pending, with Cohere expecting them by year-end. Founded in 2019, Aleph Alpha had been seen as Germany’s great hope for developing its own AI language models after the launch of the AI chatbot ChatGPT in 2022. However, the startup failed to meet high expectations and struggled to keep up with rapid technological advancements. As a result, the company shifted its focus toward building an orchestration platform for AI and industry-specific applications instead of cutting-edge fundamental AI models. Early this year, founder Jonas Andrulis left the company, possibly due to disagreements over its new direction under Scheer’s leadership. Scheer laid off 50 employees to focus more on core operations. In a September interview with the Frankfurter Allgemeine Zeitung, Scheer announced the imminent release of a new AI model. Media attention led to many customer inquiries that the company couldn’t handle due to limited capacity for such complex projects. Scheer terminated several partnerships, stating the company would no longer offer AI services unrelated to its core products. The focus is now on developing large language models and a platform to deploy them in companies and government agencies. A call for proposals for presentations at the Mastering AI conference in Karlsruhe, scheduled for April 28–29, 2027, is open until November 19.
2 reports
heise onlineIndependentCenterFactual: no official source document/info detectedObjective 70
The article announces the upcoming 'Mastering AI' conference scheduled to take place in Karlsruhe on April 28th and 29th, 2027. It mentions that the call for proposals for the specialized conference on AI projects is open until November 19th, 2023. The event appears to focus on practical applications and development of artificial intelligence projects.
Bias read (Center): The article presents factual information about an upcoming technology conference without any apparent ideological or political framing. It focuses on logistical details such as dates and submission deadlines, which are neutral in nature.
Why factuality: no official source document/info detected
Why objectivity (70): The article is highly objective as it simply presents factual details about an event without opinion, emotion, or bias. It remains neutral throughout and focuses solely on providing information without interpretation or commentary.
Aleph Alpha, once considered Germany's promising AI startup, struggled to keep up with international competitors. The company has now returned with a new language model trained specifically on German data, aiming to establish itself in the competitive AI landscape.
Bias read (Center): The article discusses a technological development by a private company and does not involve political figures, policies, or contentious issues. It focuses on the technical aspects of AI development and market competition.
Why factuality: no official source document/info detected
Why objectivity (60): The tone is neutral and informative, presenting facts about Aleph Alpha's new model without overt bias. However, the article does not provide sufficient context or balance by mentioning competing companies or alternative viewpoints.
How each side covered it
The same event, grouped by the political lean of the outlets covering it.
progressive
center
conservative
★
How each side covered it
Support independent, bias-aware news and unlock the social pulse, community voting, and every other Supporter feature.