Hacker Timesnew | past | comments | ask | show | jobs | submit | cyanf's commentslogin

I had a strange call with a support rep recently.

They sounded a tinge strange, like they’ve almost crossed the uncanny valley, only to succumb at the final 3% stretch.

I was suspicious, but their ability to understand my complex request and the relatively low latency make an LLM -> TTS or e2e voice model unlikely.

This post finally solved the mystery.


For legged, you want high torque and backdriveability (for shock absorption).

Agricultural drone motors like the eaglepower 8308 are ideal.

They’re cost effective, (~$80 from aliexpress) & you can pair them with a 3d printed cycloidal drive to fulfill both requirements.

Industry actuators are an order of magnitude more expensive than this.

Extra: If you go down this path, you’ll need a driver. The Xdrive is frequently recommended, but there’s a clone that’s significantly cheaper: https://makerbase3d.com/product/makerbase-xdrive-mini-high-p...


This is pretty cool, thanks for sharing. I wish there was a (mostly) 3d printable Cara mini, but I’ll start with Cara.


The layout is doable with hobby servos, but you'd need to patch in current sensing for that bit of the feedback. It's not terribly difficult conceptually but it's an extra complication that most servo power distribution boards don't give you.

You can also strap a capstan to the servo axle, if that's your thing. I've prototyped that myself in the past. You can go surprisingly far with an FDM printer, an SG90, and some dyneema bowstring. One thing I haven't tried is modding one for continuout rotation to get around the way the capstan drive limits the output angle you can achieve - I was happy reducing from ~180deg to ~45deg for what I was doing - but that's relatively well-trodden ground. Might pull that project out of the storage box it's languishing in at some point.


> The layout is doable with hobby servos

Mostly we're past that. Robotics with hobby servos sucks. Been there, done that, robot arm is in electronics recycling bin right now.


I agree it's not great but for scale and convenience it's good. If this was going to just get scaled down, what motors would you pick?



"The 90KV version is what you want."

You mean the "KV90 color" of course, ha ha.

That we live in an age where "agricultural drone" is even a word pairing…


thanks for the Xdrive pointer, I just got samples of tinymovr and odrive controllers (with FOC), they seem nice but are more expensive.


A seldom appreciated benefit of gruvbox: like vim binds, it's available everywhere. If it has a theming system, somebody ported gruvbox to it.


Why would you want to ssh into a machine that's not yours? That's a violation of the Computer Frauds and Abuse Act, up to 10 years in prison!


I think you're joking, but to clarify -- not personally yours. A misbehaving worker box, an app server in the staging environment, etc. A resource owned by the organization for which you work, where it would not be appropriate for you to customize it to your own liking


When you have permission to do so, it isn’t.


Tragedy of the aggregate.


There are existing solutions for queues in Postgres, notably pgmq.


Despite sentiments around Mojo being negative on HN due to the stack not being OSS, this is the ultimate goal of Modular.

https://signalsandthreads.com/why-ml-needs-a-new-programming...


I listened to that episode, by chance, last week. It was well worth the time to listen.


The blog's title can be misleading here, "we" in this context refers to the Cognition team. I don't work at Cognition, just thought this was interesting.


> On August 29, a routine load balancing change unintentionally increased the number of short-context requests routed to the 1M context servers. At the worst impacted hour on August 31, 16% of Sonnet 4 requests were affected.

Interesting, this implies that the 1M context servers performs worst at low context. Perhaps this is due to some KV cache compression, eviction or sparse attention scheme being applied on these 1M context servers?


This is due to RoPE scaling.

> All the notable open-source frameworks implement static YaRN, which means the scaling factor remains constant regardless of input length, potentially impacting performance on shorter texts. We advise adding the rope_scaling configuration only when processing long contexts is required. It is also recommended to modify the factor as needed. For example, if the typical context length for your application is 524,288 tokens, it would be better to set factor as 2.0.

https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Thinking


The key issue is that their post-mortem never explained what went wrong on two out of three issues.

All I know is that my requests can now travel along three completely different code paths, each on its own stack and tuned differently. Those optimizations can flip overnight, independent of any model-version bump—so whatever worked yesterday may already be broken today.

I really don't get the praise that they are getting for this postmortem, it only made me more annoyed.


Snappiness is the primary reason for using Zed.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: