From the Orama team
Articles
What we think about software, data and AI.
AI can now self-replicateSoftware is the one thing we know how to make that can make more of itself. AI that builds AI is already useful and measurable. What does it take to use it well, and who chooses what survives?Read the article
What does it mean to own your AI?Owning your AI isn't about owning a data center. What does a company actually need to own, what can it rent, and how do you decide which work belongs where?Read the article
What are the most valuable methods in AI?Generative AI has the attention, and it now leads the count of firms that use AI. Which methods do the most work in a business and return the most, and why do the counts disagree?Read the article
Why the AI project is central to AI successNobody can promise that an AI project will work. You can manage the ways a data project fails, one at a time. In Orama Lake, the AI project is where that is done.Read the article
Investable AI: why we built MetisMost AI in business is still a skunkworks bet, untethered from a return. Metis turns it into an investment: score the unit, check the data, design backwards from the return, and price on results.Read the article
Software is not the moatAI makes code cheap to reproduce, so code stops protecting a software business. History says cheaper software means more software, not less, and the value moves to what stays scarce.Read the article
Orama Lake is open sourceOrama Lake, the whole stack for designing, training, and shipping AI on your own data and hardware, is released under the MIT license: what's in it, how to try it, and how it's run.Read the article
