mmaunder

15 ms·

301.
by mmaunder 1y ago
When you're going nuclear on an investor who owns 49% of your company, you're in trouble.
302.
by mmaunder 1y ago
No one only uses an LLM for writing. We switch tools as needed to pull threads as they emerge. It’s like being told to explore a building without leaving a specific room.
303.
by mmaunder 1y ago
Such a cool post. Also gave me a fresh perspective on DEF CON.
304.
by mmaunder 1y ago
Yeah sorry you got downvoted, but that's pretty much my inclination is to say RTFM. Honestly though, I'm very excited by how few developers are using the most powerful tools available. Huge opportunity for the rest of us willing t
305.
by mmaunder 1y ago
Actually you can enable everything, or have fine grained control over specifics. Or you just manually approve and ask not to be prompted again. Sounds like you're more of a dabbler.
306.
by mmaunder 1y ago
Yeah it can. Either using MCP or git via bash. It’s a glaring omission and calls the data into question. How is attribution done? If it’s via the agent taking credit in commit messages, that’s a problem because Claude code, for example, has
307.
by mmaunder 1y ago
Claude 4 in agent mode is incredible. Nothing compares. But you need to have a deep technical understanding of what you’re building and how to split it into achievable milestones and build on each one. It also helps to provide it with URLs
308.
by mmaunder 1y ago
Ah thanks. Sorry I didn't see your reply before I posted the analysis. I'll leave it. Thanks for the reply. Congrats on the project. Seems like a legit need.
309.
by mmaunder 1y ago
OK well since OP isn't replying, [Edit: Author replied] it looks like they're using a wrapper process for local MCP servers and a proxy for remote, and you have to modify your MCP config to reference the local wrapper or proxy so
310.
by mmaunder 1y ago
How are you intercepting the huge variety of network calls and range of protocols that a local MCP service can make? Are you between the client and process? Or do you only support remote MCP?
311.
by mmaunder 1y ago
Probably (and unfortunately) going to need someone from Anthropic to comment on what is becoming a bit of a debacle. Someone who claims to be working on alignment at Anthropic tweeted: “If it thinks you're doing something egregiously i
312.
by mmaunder 1y ago
It’s about putting lipstick on OpenAI. They’re in trouble.
313.
by mmaunder 1y ago
Yeah Google has it all vertically integrated from the science to the chips and everything in between. It’s theirs to lose.
314.
by mmaunder 1y ago
Models are getting smaller, faster, cheaper to make, reflecting on their own output, adding modes and running in more places. But they’re not getting much smarter because they can only be as smart as us and each other, because that’s where
315.
by mmaunder 1y ago
Quote: Expanded Multimodal Understanding with Audio: Gemma 3n can understand and process audio, text, and images, and offers significantly enhanced video understanding. Its audio capabilities enable the model to perform high-quality Automat
316.
by mmaunder 1y ago
> These more abstract skills—not basic arithmetic—are essential for understanding recursion, type inference, or algorithm design. No they're not. Academia has spent decades trying to formalize many aspects of programming and continu
317.
by mmaunder 1y ago
Here's the play: Open source with AGPL, then offer an enterprise license. You get two wins. The OSS community applauds your adoption of an agressive OSS license. Enterprise customers can't use software under AGPL because it risks
318.
by mmaunder 1y ago
AI isn't replacing many jobs yet. But it is causing customers to hold off on certain purchases, like dev services, due to uncertainty. And the shills are amplifying the effect. I'm seeing layoffs caused by this.
319.
by mmaunder 1y ago
> the youth's increasing intolerance of politically incorrect ideas is caused … So, wrongthink. This is my point. And it’s incredible how the history of passing legislation to censor being ultimately used by the opposition just kee
320.
by mmaunder 1y ago
This project was. “A negative impact from the creation of Peerd is that it has created confusion among new users. I am frequently asked about the differences between Spegel and Peerd.”
321.
by mmaunder 1y ago
Great article with links to others, like this one: https://arena.org.au/stay-in-your-lane-the-oxymoron-of-authe... With quotes (re cultural appropriation) like “the ultimate endpoint of keeping our mitts off experience that
322.
by mmaunder 1y ago
If you want them to contribute back changes, use a license that makes them contribute back changes, like GPL. Don’t ever “default to” a license.
323.
by mmaunder 2y ago
Note sure why their comment was downvoted. Google the names. Hassabis runs DeepMind at Google which makes Gemini and he's quite brilliant and has an unbelievable track record. Buffet investing in teams points out that there are smart p
324.
by mmaunder 2y ago
Try uploading a bunch of PDF bank statements to notebooklm and ask it questions. Or the results of blood work. It's jaw dropping. e.g. uploaded 7 brokerage account statements as PDFs in a mess of formats and asked it to generate table
325.
by mmaunder 2y ago
This is an underrated comment. Yes it's a big advantage and probably a measurable pain point for Anthropic and OpenAI. In fact you could just do a 1% survey of robots.txt out there and get a reasonable picture. Maybe a fun project for
326.
by mmaunder 2y ago
Agreed. Once they dial in the training for sheets it's going to be incredible. I'm already using notebooklm to upload finance PDFs, then having it generate tabular data and copypasta into sheets, but it's a garage solution co
327.
by mmaunder 2y ago
More great innovation from Google. OpenAI have two major problems. The first is Google's vertically integrated chip pipeline and deep supply chain and operational knowledge when it comes to creating AI chips and putting them into produ
328.
by mmaunder 2y ago
The article tells you that. They generated a fake user accounts database. The only way to catch that is to manually spot check by contacting users. The buyer clearly didn’t.
329.
by mmaunder 2y ago
I think there are many other moving parts in this story that aren’t public knowledge.
330.
by mmaunder 2y ago
I embarked on a working holiday to the UK in 1998. Quit my job in South Africa and left with nothing but a backpack. Opened up my world, I worked for Coca Cola, the BBC, an investment bank, got wrapped up in the dot com boom, met my America

More ›