Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
ed_mercer
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
91.
▲
by
ed_mercer
8mo ago
There's no mention of RAM upgrades. If we bought RAM already at the old prices, are they being increased as well? The current pricing for RAM has more than quadrupled since January.
92.
▲
by
ed_mercer
8mo ago
> if you aren’t using AI extensively in your workflow, you are essentially trying to build a career with one hand tied behind your back. This would have been an interesting comment to make two years ago.
93.
▲
by
ed_mercer
8mo ago
Can you elaborate on “listen for changes”? AFAIK it can’t do that and needs cron jobs to check.
94.
▲
by
ed_mercer
8mo ago
Is there any developer experience on Apple products that isn’t bad?
95.
▲
by
ed_mercer
8mo ago
You can't write "autonomous agents often fail" and then advertise "AI agents that perform complex multi-step tasks autonomously" on the same site.
96.
▲
by
ed_mercer
8mo ago
Openclaw is unaffected by this as the Claude Code CLI is called directly
97.
▲
by
ed_mercer
8mo ago
But the article is from 2024! So someone at MS saw this thread?
98.
▲
by
ed_mercer
8mo ago
<3
99.
▲
by
ed_mercer
8mo ago
Missed opportunity to turn this into a poem
100.
▲
by
ed_mercer
8mo ago
We found that it works great for scraping but it's not so great for development. git worktrees in a VM are much faster over spinning up new containers with an existing codebase.
101.
▲
by
ed_mercer
8mo ago
Ahh, that makes sense. Yeah you likely have a very different business model. We're more using them as "AI employees" that help us with tasks. Currently running 4 of these VMs but planning to add more. Note that each VM has a
102.
▲
by
ed_mercer
8mo ago
As far as I'm concerned, it's already here.
103.
▲
by
ed_mercer
8mo ago
For us, we actually moved away from k8s to dedicated VMs on Proxmox for our agents. We initially had a containerized environment manager running in k8s, but found that VMs give you things containers struggle with: full desktop environments
104.
▲
by
ed_mercer
8mo ago
Looks good! I'm curious, are customers fine with their data going to third-party LLM providers?
105.
▲
by
ed_mercer
8mo ago
Did it sound like a bot from eleven labs?
106.
▲
by
ed_mercer
8mo ago
Hardware first, software second
107.
▲
by
ed_mercer
8mo ago
You can also hold ALT and press the green button to mazimze.
108.
▲
by
ed_mercer
8mo ago
OpenClaw already does all these things and better. It can run 24/7 in one (or more) VMs that I control from my phone, without extra cost.
109.
▲
by
ed_mercer
8mo ago
> Siri doesn’t always properly process queries or can take too long to handle requests, they said. Is this really a reason to delay though? It's not like the current Siri is capable of processing queries outside of "set a timer
110.
▲
by
ed_mercer
8mo ago
Is this really a valid outcome, given the facts in the article?
111.
▲
by
ed_mercer
8mo ago
Hmm. Perhaps it serves as a commitment to post about this publicly? And as one other person mentioned, you have a far better chance of beating this today than just a few years ago, especially with some money and connections. >It leads to
112.
▲
by
ed_mercer
8mo ago
You should look at some Copilot reviews.
113.
▲
by
ed_mercer
8mo ago
nods . nods again . Yep, this is exactly why we left GitHub for GitLab two years ago. Not one moment of regret. Still, I wonder who is still looking manually at CI build logs. You can use an agent to look for you, and immediately let it c
114.
▲
by
ed_mercer
8mo ago
Confirmed BS: https://resellcalendar.com/news/news/mac-mini-shortage-clawd...
115.
▲
by
ed_mercer
8mo ago
The author is a bit extreme for expecting apple to have done something as complex as ooenclaw, not even OpenAI or Anthropic have really done it yet. However this does not excuse Apple to sit with their thumbs up their asses for all these ye
116.
▲
by
ed_mercer
8mo ago
Yep, there is zero reason to use mac mini’s. It’s way more cost effective to rent one (or more!) small VMs the cloud.
117.
▲
by
ed_mercer
8mo ago
Haha yeah I noticed too late :P
118.
▲
by
ed_mercer
8mo ago
Then how do you explain the massive growth of Claude Code?
119.
▲
by
ed_mercer
9mo ago
> Microsoft itself is rolling out Claude Code to internal teams Seems like Nadella is having his Baller moment
120.
▲
by
ed_mercer
9mo ago
If you run openclaw on a spare laptop or VM and give it read only access to whatever it needs, doesn’t that eliminate most of the risk?
More ›