mmaunder

6 ms·

61.
by mmaunder 6mo ago
Bet you’re pissed.
62.
by mmaunder 6mo ago
Haven’t looked at the code, but is the server providing the client with a system prompt that it can use, which would contain fake tool definitions when this is enabled? What enables it? And why is the client still functional when it’s givin
63.
by mmaunder 6mo ago
Can we immunize HN against being yet another AI drama site? Obviously this isn’t a fundamental issue with agents or AI or Anthropic but a misconfiguration edge case.
64.
by mmaunder 6mo ago
The only barrier to a flourishing truly open source AI model ecosystem is the cost of training a highly capable model. This will get as cheap as it is to buy a computer and contribute to Linux. OSSAI movements will replace traditional OSS.
65.
by mmaunder 7mo ago
We all know they're addictive, they're designed to be addictive, and they're very, very harmful, to both adults and children. The individuals who are profiting from the harm are clearly identifiable. And that harm directly ta
66.
by mmaunder 7mo ago
We're calling agents harnesses now?
67.
by mmaunder 7mo ago
I’d encourage devs to use MiniMax, Kimi, etc for real world tasks that require intelligence. The down sides emerge pretty fast: much higher reasoning token use, slower outputs, and degradation that is palpable. Sadly, you do get what you pa
68.
by mmaunder 7mo ago
I'm a ground instructor and instrument rated pilot and I fly a 206 in and out of busy charlie and delta airports. I'm also a ham radio guy (WT1J) and an SDR dev. I'm 100% with you on this, but the amount of inertia you'r
69.
by mmaunder 7mo ago
Yeah just try to select text to copy. Nope. Try to scroll back in terminal or tmux. Nope. Overbearing for sure.
70.
by mmaunder 7mo ago
Given the free market nature of cellphones, where vendors and companies have unfettered access to monetize users, having cellphones in school is akin to making school children line up and listen to sales pitches from companies around the wo
71.
by mmaunder 7mo ago
This feels like a response to openclaw (and openai's hiring of the lead).
72.
by mmaunder 7mo ago
Almost too good to be true. They didn't find large quantities of weed, and afroman had cameras set up and caught it on camera. I mean, talk about landing with your bum in the butter. His career just caught a major reboot.
73.
by mmaunder 7mo ago
I think she got a few "oh shit she spoke truth" reactions around her which is what prompted that.
74.
by mmaunder 7mo ago
Have you tried it? It's like working with an idiot savant. It's absolutely brilliant, but goes off the rails constantly, spewing out CoT when it shouldn't be, getting into weird loops, spewing gibberish or repeated phrases. B
75.
by mmaunder 7mo ago
Yes. CLI. Always CLI. Never MCP. Ever. You’re welcome.
76.
by mmaunder 7mo ago
Google is so far behind agentic cli coding. Gemini CLI is awful. So bad in fact that it’s clear none of their team use it. Also MCP is very obviously dead, as any of us doing heavy agentic coding know. Why permanently sacrifice that chunk o
77.
by mmaunder 7mo ago
Cheapest four letter domain on Earth at this point, given the negative value of the business and brand.
78.
by mmaunder 7mo ago
OP can you please make it not as dark and slightly larger. Super useful otherwise. Qwen 3.5 9B is going to get a lot of love out of this.
79.
by mmaunder 7mo ago
That's between 1 and 10 training runs on a large foundational model, depending on pricing discounts and how much they manage to optimize it. I priced this out last night on AWS, which is admittedly expensive, but models have also gotte
80.
by mmaunder 7mo ago
Someone explain how you'd create a vector embedding using homomorphically encrypted data, without decrypting it. Seems like a catch 22. You don't get to know the semantic meaning, but need the semantic meaning to position it in hi
81.
by mmaunder 7mo ago
Nice. I played with this a bit. Agents are very good at Rust and CUDA so massive parallelization of compute for things like options chains may give you an edge. Also, you may find you have a hard time getting very low latency connection - o
82.
by mmaunder 7mo ago
That's one way to block those pesky young innovators from trampling our lawn.
83.
by mmaunder 7mo ago
Same. 52 year old CTO here.
84.
by mmaunder 7mo ago
Ah - I know! Seriously I know. There's such a bad need for this right now. The problem is that the folks who are great at agentic coding are coding their asses off 16 to 20 hours a day and don't have a minute they want to spend on
85.
by mmaunder 7mo ago
I kind of agree. But I'd adjust that to say that in both cases you get good looking code. In the hands of a junior you get crappy architecture decisions and complete failure to manage complexity which results in the inevitable reddit &
86.
by mmaunder 7mo ago
That's actually a great question. Truth be told the best way right now is to grab Codex CLI or Claude CLI (I strongly prefer Codex, but Claude has its fans), and just start. Immediately. Then go hard for a few months and you'll de
87.
by mmaunder 7mo ago
Agreed 100%. I'd add that it's the knowledge of architecture and scaling that you got from writing all that good code, shipping it, and then having to scale it. It gives you the vocabulary and broad and deep knowledge base to inno
88.
by mmaunder 7mo ago
But my AI didn't do what your AI did. Cherry picked AI fail for upvotes. Which you’ll get plenty of here an on Reddit from those too lazy to go and take a look for themselves. Using Codex or Claude to write and optimize high performanc
89.
by mmaunder 7mo ago
Weird and frustrating that this has not hit the HN HP. I'm not the poster. Huge development in cybersec.
90.
by mmaunder 7mo ago
Google really know how to screw up a product experience. npm install -g @googleworkspace/cli gws auth setup { "error": { "code": 400, "message": "gcloud CLI not found. Install it from https

More ›