In the News

• Susty Ventures gets its’ first media mention in TechCrunch! This feature comes at a time where Susty is working with some great founders as in the article, and others, including in healthcare. The focus in short is “sustainability on the ground”, in particular how new ventures can structure themselves from Day 1 around this opportunity, and win in the long term because they look at problems differently.

At a time when many global signals are very much in no good territory, a silver lining to ponder on is that previously steadfast assumptions about how the world works are being challenged out in the open in real-time, shifting perceptions and requiring a response, not least of which supply chains which are (were) hyper optimised and are now having to adapt if only to pervasive uncertainty. However, this could very well be a wrenching possibly extinction-level disruption in the short term, difficult to weather and imagine that there is much leftover optimism when this adaptation period ends.

AI looks to be a harder integration sell as a result - it takes a concerted, sustained effort to get value out of most AI at any scale greater or more general than a few power-users, whether in development with engineering tasks, like fine-tuning, or in production, like the necessary prompting back-and-forth to get the expected results. One way to pave this challenge over is to just mandate AI usage top-down, but this risks lasting only as long as confidence in the model while new ones get released.

On the end-user side (which many buyers will end up being), AI thus is quite “leaky” - prompts get retrained, new models come out, and what previously appeared to work suddenly doesn’t, with no indication why that is the case (since the AI entity itself is not aware of the problem)

• MCP and the rest of the LangChain follow-ons are attempts at futhering AI tooling, but a lot of important tasks in corporate are not suited to that level of hands-off piloting. Even so, they are being boosted with vibe-coding chatter, the latest passing fad of signalling that one is jumping into a project without a very clear idea of what to build nor how to manage it, wholly encouraged by an often uber-chipper LLM.

In the physical space, driverless cars are gaining traction in a few US metros, but it still doesn’t quite make sense to have a common good like this exist peri-municipally - but banging on the drum of civics is overplayed in a world where only too few externalities are priced in. Still, accounting for these and building them into the business model might allow entrants to establish their position more effectively as a result, much like highways tolls earlier and congestion pricing today being such a stable revenue stream that cities were generally content to fund them by way of collaterized bonds.

There still are many ways to succeed, the technology is still changing, some problems are thrown into sharper relief now, and collaboration around a shared future is as important as ever.