Joined June 2007
954 Photos and videos
luma.com/e0cuhtx8 Is your AI enterprise running day and night? It should. Are you having hundreds or thousands of employees in a collaborating PPOacting AI infrastructure? You should Allow us to show you in a two-day, high-impact program designed for executives and senior decision-makers. #WorkflowIntelligence, #OperationalExcellence, #ExponentialProductivity, #EnterpriseAI, #BlueCallom·AI #BlueCallom
1
19
Check out the latest article in my newsletter: Internship - The New MBA for AI Careers of any age and any kind. linkedin.com/pulse/internshi… via @LinkedIn
29
How do Enterprise AI Applications generate the most value? The most important aspect to consider: * Existing applications on new platforms don't change the world. * New applications, which cannot be built on existing platforms, make all the difference. *How AI solves the energy problem it created* ---------------------------------------------------------- Energy is scarce, and when the energy-hungry AI shows up, it's like a Wildebeest arrives at a waterhole, where birds, antelopes, and even lions make way. Interestingly enough, nobody in the AI Space used AI itself to solve the problem. A Swiss AI firm with a unique neuroscience-inspired AI Innovation System did. The result stunned even scientists. Geothermal Energy like you never thought of. It goes 35km deep into Earth's crust, where temperatures are comparable to those in nuclear power plants: 500 °C. Modern Physics, however, was in its way. The deepest hole ever drilled was 12 km. Now, China is conducting an experiment with lasers to get to 14 km, but it is still far from the much-needed 20 km. The solution is so unbelievably simple that everybody must shake their head. AI can create the energy crises it creates with it's own technology. An AI, trained to be disruptive BlueCallom DISRUPT, an agentic AI System, made a move similar to move 37 at Alpha GO - nobody was expecting it. In our example, instead of a vertical hole that goes 20 km deep, it bores a tunnel 35 km long at a 45° angle. The Steepest cogwheel train in Switzerland is at 47 degrees, and the longest tunnel, also made in Switzerland, is 50km long. Cracking the impossible Months before: After countless unsuccessful attempts, we were almost giving up. The claim that it has access only to standard information and cannot generate new ideas was still ringing in my ears. But we solved smaller problems that were also nowhere to be found, and the AI literally created new ideas. As we considered differences in thinking between highly creative and less creative people, we realized that people like me were trained to think outside the box. Thanks to Neuroscientist David Eagleman, I learned how the brain composes thoughts, then extended that concept to ideas. Which is generally the same process but with different inputs. So we 'INSTRUCTED' the AI to think very differently. It took us a week to get to stunning results. The biggest learnings: This is not about the LLM, - it is about how we instruct the LLM It is not only about generating disruptive ideas; we also need to develop a plan to make them a reality and execute them. 10 years ago, I wrote: "The initial Value of an Idea is exactly Zero" Making it feasible The AI application has now conducted a feasibility study. One angularly bored hole could deliver 1.6 GW; 1000 holes generate 1.6 TW. Earth's crust has vast geothermal potential (estimates suggest that just 1% could yield 63 TW, far exceeding today's ~30 TW global electricity demand). Nearly 450 Reactors globally produce approximately 1 TW. Angbor concept could quadruple electricity generation, without plutonium, coal, wind, solar, or water. And the early price calculations show that capital expenditure could drop from €12 Bn for a nuclear power plant to 7 Bn per Geothermal plant. The operating cost (opex) could decrease by 25%, primarily due to reduced security requirements. Especially countries with limited natural resources, such as plutonium, or with large water reservoirs, could benefit from the Angbor concept. Innovation is just one of many examples of how AI Applications change the world. Applying this thinking shows how Enterprise AI can change Innovation in an instant. The above concept was developed over one weekend, including a 250-page report with a blueprint for how to build it, comprising almost 100 pages. Doing it yourselves will cost approximately $/€/CHF 5,000, save your team 6 months, and reduce your time-to-market. That's an instant ROI. I will continue this series and talk about other AI Applications that would be impossible to build with conventional software or prohibitively expensive. By the end of this year, we will have around 30 of those applications. Competition Management --- An application that you can run weekly, which observes your entire market, relevant technology trends, your competitors, and other key factors. Sales Management --- An application that navigates your sales people through customer and case-specific sales workflows, provides situation-specific suggestions, provides reminders or inspirations, and increases customer face time to 80%. Forecast, reporting, funnel control is all done by the AI Product Launch, Compliance, Webinar Management, Talent acquisition, and Logistics all have one thing in common: they are either not even available in the old world or profoundly different. All are integrated with one another and designed for entire teams rather than individual users. You can subscribe to the "Enterprise AI Solutions" newsletter to avoid missing any of the Enterprise AI applications, their unique technologies, and the platform evolution on which they run. How we did it: bluecallom.com/bluecallom-di… Where can you learn to turn your company into a Cognitive Enterprise? bluecallom.com/enterprise-ai…
34
Axel Schultze retweeted
DeepSeek-R1’s paper was updated and expanded from 22 pages to 86 pages two days ago The R1 release anniversary is coming up on January 20, could we hear from the whale soon? Another interesting date is February 17 (Chinese New Year), as DeepSeek often releases updates around Chinese holidays
DeepSeek-R1’s paper was updated 2 days ago, expanding from 22 pages to 86 pages and adding a substantial amount of detail. The new content covers topics such as the self-evolution of DeepSeek-R1-Zero, evaluation of DeepSeek-R1, further analysis, and DeepSeek-R1 distillation. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning Paper: arxiv.org/abs/2501.12948v2
12
33
388
38,588
21 Dec 2025
Check out the latest article in my newsletter: 2026 - Jahr des Enterprise AI. linkedin.com/pulse/2026-jahr… via @LinkedIn Many large businesses died in the past due to a lack of adaptation. Let's see who learned from this.
35
9 Oct 2025
Ich bin kein Politiker, aber wirklich bestürzt, dass man über so ein unglaubliches 2.000-Seiten-Dokument in der SCHWEIZ überhaupt nachdenken muss. Dieser Vertrag MUSS GESTOPPT WERDEN. unterwerfungsvertrag.ch/ #Unterwerfungsvertrag #SouveräneSchweiz #neutralität wieder herstellen #unabhängigkeitstag
1
38