does feel like something from the RAG era
@petergostev
-
Mythos monitors your open source repositories automatically
By
–
Mythos is watching your open source repos while you sleep
-
Anthropic acquires new building for expansion
By
–
Wow can't believe Anthropic snapped up this building
-
Agents AI autonomous workflows eliminating manual human UI interactions
By
–
I am getting radicalised against human only UIs. Let my agents cook and don't make me press another button ever again.
-
Engineering Culture: The Cost of Shipping Without Maintenance
By
–
A common problem we will now see is the kind of 'google disease' but at mass scale. Where smart engineers can ship apps quickly, then lose interest and move on to something else cool to ship. It feels good to ship a whole new app, but less cool to keep fixing bugs a year later. https://t.co/thFu2O8T2N
— Peter Gostev (@petergostev) 16 avril 2026A common problem we will now see is the kind of 'google disease' but at mass scale. Where smart engineers can ship apps quickly, then lose interest and move on to something else cool to ship. It feels good to ship a whole new app, but less cool to keep fixing bugs a year later.
-
Challenges of Implementing AI Code in Enterprise Tools
By
–
I don't feel like people would be vibe coding slack or workday, it is still a lot of work getting these right even with perfect coding ability
-
Running shoe company pivots to AI data centre operations strategy
By
–
Why is everyone freaking out? Pivoting from running shoes to running AI data centres is such an obvious idea
-
Building Agent-Friendly App Versions as Startup Arbitrage
By
–
A good start-up idea is to take any popular app and make an extremely agent-friendly version of it – sort of an 'agent arbitrage'. Slack, Workday, Notion, Figma, Navan etc. are all very clunky with agents, even if they have an agent/MCP, they are nowhere good enough. I am now
-
Users Frustrated with Current AI Models Demand Next Generation
By
–
I know it's stupid, but I've had enough of the sharp edges of the current generation of models. Bring on the next one. I want to feel something.
-
Reinforcement Learning Speed Trade-offs in Large Language Models
By
–
Significant downside of larger models is that it is much harder to do RL on them. Smaller model = quicker RL cycles, bigger model = slower RL cycles. So far this matches – GPT-5 is the smallest model with quickest iterations, Gemini 3 is biggest and slowest iterations, Claude is