DeepSeek V4.1 Flash is now on Verda Inference, at about 50% lower cost.*
@deepseek_ai's capable, efficient model for everyday coding and agents, with a 1M-token context and image input.
→ $0.15 per 1M input tokens, $0.60 per 1M output
→ Cached input at $0.006
→ Thinking on or off
Read the announcement: https://t.co/67irjVRPoj
*Based on published list prices of most providers offering DeepSeek V4.1 Flash, as of Nov 13, 2026. Prices are subject to change.
I just fixed my dishwasher with the help of ChatGPT. A trivial task. I had been about to order a new one. So this software has increased the country's real wealth yet decreased the measured GDP. The main economic indicator is structurally incapable of registering the thing that actually makes people better off, namely the growth of knowledge
I discussed with a Senior Engineer at work about ai agents. He uses them a lot and wondered why I'm still writing code by hand during my free time. I told him that last weekend I deepened my understanding of Unix sockets thanks to coding by hand.
He said "pff well yeah of course you have to know the fundamentals to know that the agent is doing something good or not".
Us junior engineers haven't yet learned all the fundamentals. Senior engineers have. This is why they are better at using agents: ability to code predicts your ability to guide agents.
So junior engineers: don't listen to everything the experienced engineers are saying. If you want to continue in this field, you have to understand the fundamentals, deeply. And you won't gain that understanding by just coding with agents.
@akoskm For us junior programmer's it's very difficult to balance learning vs producing output. It's hard to learn computer science with just agentic coding. It's harder for us to guide the agent.