If you're trying to persuade someone to try Mastodon (or Pixelfed, or Friendica, etc), you'll get a lot further explaining them as algorithm-free, ad-free, AI-free, nazi-free conversations, rather than trying to describe a "decentralised protocol" or why "open source" is good. Mastodon is conversations between real people. It's sharing.

Deploy your AI assistant in minutes. Telegram ready, WhatsApp coming soon.

What does agentic coding mean for software as a business model? I see two fundamentally different types of software: software that is used by humans, and software that runs on its own. The impact will be very different. (𝟭) 𝗛𝘂𝗺𝗮𝗻-𝗢𝗽𝗲𝗿𝗮𝘁𝗲𝗱 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 (SaaS, desktop software, mobile apps, ...) Thanks to coding agents, developers have become dramatically more productive (and we are still in the early days). As a result, the amount of software will explode: more new apps and far more features in existing ones. But there is one thing that is not growing: human attention. Neither the number of users nor the available time per human will increase at the same pace. This leads to a massive oversupply of software. In other words, humans simply do not have enough time to use all that shiny software. As a result, the audience per piece of software will shrink. Today, a successful B2B SaaS can be used by hundreds of thousands of people or more. With rising competition, this number will inevitably go down. At the same time, users will expect increasingly tailored solutions. In the past, customer-specific software was a big no-go in B2B. Going forward, it will be the default expectation. It is likely that a lot of software will be used by only one customer: a single company, or even a single human. There is even already a term: Service-as-Software. Obviously, this will have profound effects on the software industry. Development teams will shrink because traditional setups were built under very different constraints. Often, a single skilled developer will be enough to build and run software end to end. Service providers will struggle to sell large teams. Software vendors will have to rethink their business models and pricing strategies from the ground up, because the underlying assumptions are no longer stable. What still matters are the classic software moats: network effects, lock-in, access to proprietary data, financial flywheels, and hardware. The large players are very aware of this. That is why we are currently watching multiple wars unfold at the same time: AI browsers, coding agents, AI assistants, and more. (𝟮) 𝗔𝘂𝘁𝗼𝗻𝗼𝗺𝗼𝘂𝘀 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 There is a second type of software that is not limited by human attention: fully autonomous agents. Most people are unaware of the high-speed tsunami wave that is about to change their lives. This is where I want to place my personal bets. It is an effectively infinite market, and we are still at a very early stage. There is a lot to discover, and today we are facing many constraints. In the long term, however, this is what will push humanity to the next level. | 16 comments on LinkedIn

Remembers everything, understands you entirely. Pickle OS seamlessly collects your context from every source and organizes it into a single cluster for infinite memory.

Entdecke überholtes Wissen aus der Schule – Fakten, die früher als wahr galten, heute aber widerlegt sind.