Of course, the latest DeepSeek models are not as good as Claude, but they're not super far off either.
The risk of letting your agent read .env goes far beyond the risk that the agent itself does something you don’t like with the contents.
Gitlab is going to take off? This is not investment advice.
Even acknowledging we don't know exactly what costs would look like in a world without VC money, wouldn't hosting models logically be cheaper to do at scale in a data center?
When I compared to the cost of running DeepSeek locally, I meant that we can treat that cost as a price ceiling, not the floor.
No, I think local stuff using also-useful-for-other-things hardware will vastly undercut cloud hosting when the free money pipeline shuts down, and will stay that way for roughly forever. That doesn't mean cloud stuff isn't useful, clearly it is, but adding another company in the middle is rarely the solution for reducing costs.
It's especially a crazy assumption to make relative to the costs of employing a human. The costs of paying an entry level employee are unlikely to go down at all, and even if those costs do decline, there's a floor they can't drop below (minimum wage at the extreme end), whereas companies are free to optimize agentic costs as close to zero as possible.
So you are assuming that a cost which is extremely susceptible to optimization but which no one has yet seriously attempted to minimize will remain perpetually above a cost which is much less susceptible to optimization, is already subject to enormous efforts to minimize, and has a legally mandated floor. That seems like a bad bet.
I’ve spent $10-$20 a day using Claude to write code and closer to $5 a day now that I mostly use Deepseek and GLM, using API pricing (no subscriptions) since I don’t use Claude Code.
This is a rounding error for a company. So I think there’s plenty of room to use AI extensively while being more cost-conscious.
Agents are expensive in large part because tool calls require round trips. It's because these APIs are stateless and not streaming so you have to resend the whole context each time. This means you have roughly #tool calls x 1/2 context size cached input tokens over any given session. Most API providers overcharge you by a huge amount for cached tokens. A exception being Deepseek. Paying OpenAI $0.05 for 100k cached GPT5.5 tokens during a possibly 2 second round trip agent tool call is like paying $100/hr for what is likely to be ~10 to 20 GB of VRAM residence (holding the KV cache).
Or it got offloaded to NVME and you are paying $0.05 for that much PCIe bandwidth.
I'd imagine GPT-5.5 and Claude Opus 4.7 could run just fine on a 16x H200 node and serve at least 10 heavy users without the token output getting choppy.
The financials don’t make sense now. Based on the expenditure the finances won’t ever make sense.
I also don't think that blitz scaling will work like with Uber. The engineers are still there. We can work without the LLM tools.
The world will look drastically different 5 years from now; for the better or worse, so save every penny (especially if you work in tech).