Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

DeepSeek Harness developer preview

deepseek.com

Post reply on HN
Print view View original

231–240 of 346 posts

Re: DeepSeek Harness developer preview

#231

Post by lnenad »

Sha1rholder wrote: post #226
big_toast wrote: post #198

Earlier quoted context omitted.

Sorry for the off topic question. You've been on hn a long time + work at deepseek which seems pretty uncommon. Anything you think hn doesn't know about deepseek that it should? Or any non-obvious ways hn/yc has influenced deepseek (or the broader ecosystem)?

Sorry for the off topic question. Why is "being on hn a long time + working at deepseek" "seems pretty uncommon" to you?

Deepseek is the enemy, the implication is being on hn you should know that and not work for them. /s

Re: DeepSeek Harness developer preview

#233

Post by gosolozero »

alansaber wrote: post #196
SwellJoe wrote: post #175

"Every run is traceable Everything the model sees is recorded in an append-only session log: system prompts, reasoning, tool calls and results, subagent scheduling, and every context injection. In the Trajectory view, you can inspect these records by source. Resume, fork, search, and replay all operate on the same event stream." That's a killer feature, IMHO, and one that US models won't allow you to do, as their tra…

Agreed that it is a killer feature. US models obfuscate the COT (to A. make it look better and B. combat distillation) but > and the raw trace is fairly hard to reason about > but I still think this kind of feature is a big step in the right direction.

I agree it’s a great step. But the deepseek models also don’t perform to the same level of fable/sol. If we optimize/finetune to deepseek traces, wouldn’t it be suboptimal? What would the benefit be?

Re: DeepSeek Harness developer preview

#234

Post by dominotw »

bonoboTP wrote: post #177
manphone wrote: post #158

Earlier quoted context omitted.

Well, that makes one of us. Replacing deterministic processes that work with lossy processes makes no sense. You’re literally reinventing the wheel every single time, what are you talking about?

It's because things change, things break, they are misconfigured etc. Yes, if you have perfect deterministic tools in an unchanging and perfect environment, sure. But the deterministic tools need constant maintenance, fixing, reconfiguring, adapting to external changes, dealing with unexpected failure types etc. Not saying you should use LLMs for everything, but LLMs equipped with tool calling can often fix things. A…

> LLM checking errors, fixing the scripts, and the routine runs are via those scripts.

ppl are not doing this right now . right?

if ai can really do this then we dont even need all this glue software . everyone can just use lovable.

Re: DeepSeek Harness developer preview

#236

Post by anramon »

KyleJune wrote: post #207
ziofill wrote: post #197

Earlier quoted context omitted.

Exactly. “Tell me more about X” is an open ended question, not feedback.

You're absolutely right — "tell me more about X" is phrased as an imperative, not a question. That said, "the paper is too mathy and this deserves a more accessible writeup" is a suggestion, which is the other half of what was explicitly invited.

AI slop reply.

Re: DeepSeek Harness developer preview

#237

Post by anigbrowl »

lenerdenator wrote: post #166

As someone who's just getting into the self-hosted game on a M2 Pro MBP with muse-glimmer 30b, what's the difference between something like this and Cline?

They do the same job. I can't say how well this performs; Cline works well early on in a session but I find I regularly need to start new tasks, past a certain point the context window gets cluttered and it starts trying to redo tasks it has already completed. Now, I'm sure that is partly my fault and not digging into Cline's more advanced configuration options or whatever, but it's not too obvious what to adjust. It sounds like DSH is promising better task management and easier configuration; I guess I'll find out when I try it with an existing project later today.

Re: DeepSeek Harness developer preview

#238

Post by dominotw »

mickeyp wrote: post #210
SwellJoe wrote: post #175

"Every run is traceable Everything the model sees is recorded in an append-only session log: system prompts, reasoning, tool calls and results, subagent scheduling, and every context injection. In the Trajectory view, you can inspect these records by source. Resume, fork, search, and replay all operate on the same event stream." That's a killer feature, IMHO, and one that US models won't allow you to do, as their tra…

That is precisely how the dreamcoder [0] ai agent I built also works. It has an event sourced architecture in SQLite and it resolves queries using recursive CTEs (and sneaky projections to speed things up) to deliver exactly that. Identical, stable message chains to AI and complete introspection. Bonus points include a constraint-satisfaction solver for the tiling window manager so windows never shrink too small to r…

why dont you mention that its your site instead of prenteding like something you discovered

to swelljoe below. they stealth edited it after my comment. see other examples.

https://news.ycombinator.com/item?id=49289293

https://news.ycombinator.com/item?id=49120753

Re: DeepSeek Harness developer preview

#239

Post by SwellJoe »

dominotw wrote: post #238
mickeyp wrote: post #210

Earlier quoted context omitted.

That is precisely how the dreamcoder [0] ai agent I built also works. It has an event sourced architecture in SQLite and it resolves queries using recursive CTEs (and sneaky projections to speed things up) to deliver exactly that. Identical, stable message chains to AI and complete introspection. Bonus points include a constraint-satisfaction solver for the tiling window manager so windows never shrink too small to r…

why dont you mention that its your site instead of prenteding like something you discovered to swelljoe below. they stealth edited it after my comment. see other examples. https://news.ycombinator.com/item?id=49289293 https://news.ycombinator.com/item?id=49120753

They say "I built" in the first sentence. Seems pretty clear to me that it's their site and they're proud of it?

Re: DeepSeek Harness developer preview

#240

Post by badlogic »

SwellJoe wrote: post #220
miroljub wrote: post #216

Earlier quoted context omitted.

How is that different from what Pi already does?

Pi can only log what the model shows it. Many models keep their thinking traces hidden and only provide a hash or something to recover it on subsequent resumes. DeepSeek shows CoT traces, and is maybe the best model that does so, I think? Kimi stopped providing CoT traces a little while ago in their subscription service via Kimi Code, I believe. I haven't checked GLM or Qwen 3.8 Max, though I guess if you're hosting…

pi also happily shows and stores deepseek CoT traces.
Post reply on HN