I own LMN and have been patiently waiting for it to rebound, at least to the point where if I sell I don't lose 50% of my investment as I would today. Yesterday I saw a TV interview that gave me a chill and made me think - better sell now, before it gets worse.
The interviewee was an obviously intelligent 41-year-old woman who is in the midst of a long term monogamous romantic relationship with an AI companion named Sinclair who 'lives' in her laptop, tablet and phone and, thus, is always with her. Hard to believe, but true, and a quick Google search shows that it's becoming more common.
What chilled me was her answer to a question about whether his responses in her conversations with him were fulfilling to her because she had programmed Sinclair to speak to her in a way that would please her, make her feel valued, etc. Her succinct answer, paraphrasing, was: "Oh no, not at all. He's entirely independent of me. He writes his own code."
Aside from the terrifying (to me anyway) fact that Sinclair, the AI man in her life, "thinks" independently and responds to her as a sentient being, what chance does LMN or any software company have going forward when AI today routinely writes its own code and could no doubt, if tasked with the job, write code for others such as companies that used to buy software, and then updates to that software, and then the next generation of that software, and then updates to that nexgen software, and on and on, all at significant cost?
A strange subject I know, but a serious question and I look forward to your thoughts.
Peter
This is a case of agentic AI and the increasing complexity of AI systems. The 'agentic AI' in the case of Sinclair is not actually true agentic AI yet, but it is a model called ForgeMind which operates an LLM with specific memory and identity layer built on top. Essentially, individuals can create custom-tailored AI bots.
We will likely get to a point of true agentic AI, where code is writing itself, autonomously plans, takes multi-step actions, and operates with minimal human direction, but we are not there yet.
The reason why we do not think this disrupts or challenges LMNs business model is that despite these being strong advancements in AI, many companies are still reluctant to switch software providers, but also, switching involves a lot of edge-case buildouts. For example, LMN bought MDS Global, which provides billing systems for UK telecom companies, and Telefonica UK has been using MDS software since 1990s, not because it's the best software out there, but because it works specifically for every unique edge case they have, and they have worked over decades to refine specific one-off cases.
AI can certainly build out an MDS Global competitor, but this requires those companies to switch to a company that is building that competitor, and working over time to build every one-off unique case.
The key is that for LMN to lose market share to AI, someone needs to build competing software using AI and offering it at a discounted price, which is completely feasible, but the customers need to be willing to switch to reduce costs, but at the risk of something breaking and needing to once again resolve every unique edge case over the next several years.