Adobe's moat may be that no artists want to build their own Photoshop or Premiere, even as everyone in finance thinks they can vibe code Bloomberg and has the money to try. x.com 

Figma faces a long-term competitive threat because its users have already shown they will switch: every Figma user left the Adobe suite once a better option emerged. If a genuinely better suite at a similar price came along, from Anthropic for example, they could switch again, enterprise layer included. A post the author replies to argues Figma's moat is its governance layer and system of record, citing 46% growth and 139% net dollar retention. A short-term software rebound remains possible. … Show more

Claude can replace Adobe for PDF editing in minutes. A custom editor took 15 minutes to build, avoided paying Adobe and had more options; Adobe is done for the long term. x.com 

Duolingo's accelerating daily user growth shows its durability and suggests fears of AI killing education and translation are overblown. Q2 2026 DAUs grew 23% year over year to 58.7M, up from 21% growth in Q1. x.com 

Duolingo is not worth owning because its core product can be replicated by AI, and the market compresses the multiple before revenue falls once a moat is gone. It has 150 million-plus monthly active users with ingrained daily habits and is integrating AI faster than most, making it more defensible than average software, yet it is still down 80% this year. Software companies with replicable core products are being rerated lower across the board. x.com 

Adobe is hard to want to own after an open-source project reimplemented its full software suite, as @math_fournity reports. x.com 

Duolingo's teaching know-how is harder for generic AI bots to copy than language knowledge alone. Its designers teach recognition of the literary French tense passé simple through La Fontaine's fables, without encouraging its use in everyday speech. Years of testing lessons on millions of learners cannot simply be scraped from the internet. AI-disruption bears seriously underestimate that know-how, which should become more valuable as AI gets cheaper. x.com 

Duolingo faces bottom-up disruption as agentic AI and large language models build tutoring in: anyone can ask them to teach a topic, make flash cards or prepare for a business trip, leaving too many surfaces to attack it. @StockSavvyShay's view that a roughly 75% drop and under-15x earnings may already price in the fear is not going to end well. x.com  x.com 

Adobe faces an existential AI question: if anyone can edit an image or generate a video with a text prompt, what is Photoshop for? Image and video generation is changing faster than almost any field but code, threatening the complex interfaces Creative Cloud was built on, and Adobe relies on Google, OpenAI, Midjourney and others for frontier models. Canva and startups are attacking pieces of the suite. Pixel-perfect editing and Firefly's $250M ARR, up 75% a quarter, are the bull case. x.com 

Duolingo still offers the best learning experience: trying it to learn Spanish turned up nothing else as good. Rivals can be built quickly and the free version is annoying, but it works. At $104 the stock is testing its breakout level at the 50-day moving average, a decent-looking setup. x.com