sanderjd

11 ms·

361.
by sanderjd 6mo ago
Yeah I relate to this. I think working in smaller chunks helps a lot. (Just like how it is for work done by humans!)
362.
by sanderjd 6mo ago
I really don't relate to this... The way you apply the expert loop is to be the expert. "Can we try this...", "have you checked that...", "but what about...". To some degree you can try to get agents to wo
363.
by sanderjd 6mo ago
Aren't they more likely to leave?
364.
by sanderjd 6mo ago
Yeah but does this work? Are there companies doing this successfully?
365.
by sanderjd 6mo ago
Which workplaces?
366.
by sanderjd 6mo ago
This is true. But I find AI tools to be a huge help for all of this. Not to do any of it faster , but to remove a bunch of the tedium from the process of testing ideas and iterating on them. Instead of "I wonder if the problem is...&q
367.
by sanderjd 6mo ago
Yep, validation is key. The smartest thing I've heard on this, which has reoriented how I think about this is that the objective function of a piece of software is now more important to get right than the implementation.
368.
by sanderjd 6mo ago
Hmmm, I think I disagree with this. I estimate that I'm now spending about 10 to 30 hours less time a week in the mechanical parts of writing and refactoring code, researching how to plumb components together, and doing "figure ou
369.
by sanderjd 6mo ago
I think if companies feel that AI usage turns out to be wasting money, with negative ROI, then certainly nobody has anything to worry about here! Companies will definitely turn off this spigot the moment they think it's a net negative
370.
by sanderjd 6mo ago
There are rules against using performance enhancing drugs in competitive sports, because ultimately the goal of the sport is entertainment, and the entertainment value increases with a rule based even playing field. Business is not like thi
371.
by sanderjd 6mo ago
No, see, this is the disconnect. Whatever happens with this in the future is not due to "the more profound consequences of [my] actions". Whether I choose to take advantage of these useful tools, or not, has absolutely no bearing
372.
by sanderjd 6mo ago
The total opposite thing :)
373.
by sanderjd 6mo ago
1. Communicate with people about what to build, fix, or improve, and about our progress with current ongoing work. 2. Figure out what components already exist and what new things we need to build and how things should be integrated. 3. Actu
374.
by sanderjd 6mo ago
Yep I understand this. But what I don't understand is, "wow this is incredibly useful! ... but I'm not gonna use it".
375.
by sanderjd 6mo ago
I understand and share your concerns but (without thinking I'm such a good engineer that companies will kill for me), I just don't share your conclusion. To me, it's pretty simple. I have things to do. This makes it easier fo
376.
by sanderjd 6mo ago
Yep, I'm a fan of the current crop of AI tools because they are incredibly useful to me, but I have deep concerns about the pipeline problem.
377.
by sanderjd 6mo ago
Seems like the upside is that it makes the job way easier? What am I missing?
378.
by sanderjd 6mo ago
Only 45x?
379.
by sanderjd 6mo ago
Most of the discussion here is about anthropomorphizing, which I honestly think is a bit of a distraction. The third one about responsibility is the most important one, IMO. This was attributed to an IBM manual decades ago, and I think it r
380.
by sanderjd 6mo ago
This is probably right, but just to note that it's very much a generational thing. When I got a TI-83 (and then eventually an 89!) it was easily the most advanced handheld computing hardware I had ever been exposed to. The iPhone made
381.
by sanderjd 6mo ago
Which open weights models did you use for this comparison, and how are you running them?
382.
by sanderjd 7mo ago
I'm beginning to develop the opinion that the next step in this process will (or at least should) be local and/or self-hosted inference. The latest qwen models are already very useful, and the smaller ones can be run locally on my
383.
by sanderjd 7mo ago
Well said. Yeah. Well intentioned things can still result in bad outcomes.
384.
by sanderjd 7mo ago
It's because none of the stuff you say is obvious is actually obvious. You might be totally right about all of it (my own view is that regardless of what the intention is, this stuff will inevitably be misused), but it needs to be demo
385.
by sanderjd 7mo ago
I kinda figured this out from context clues, but they really should introduce the term on this page!
386.
by sanderjd 7mo ago
Productizing openclaw. Seems pretty smart. (But probably going to be commoditized pretty quickly.)
387.
by sanderjd 7mo ago
Right, we truly don't have a strong rule about differentiating these in the standard American dialect! Most people say STINE for this one, but if you say STEEN, nobody is gonna be confused or tell you that it's wrong.
388.
by sanderjd 7mo ago
Thanks! This is very neat. BTW, I went to your website looking for this, but didn't find your blog. I do now see that it's linked in the footer, but I was looking for it in the hamburger menu.
389.
by sanderjd 7mo ago
Maybe! Or maybe there is really a competitive advantage to "artisanal" coding. Personally, I would not currently expect a fork of RedoxOS that is AI-implemented to become more popular than RedoxOS itself.
390.
by sanderjd 7mo ago
As someone who enjoys working with AI tools, I honestly think the best approach here might be bifurcation. Start new projects using LLM tools, or maybe fork projects where that is acceptable. Don't force the volunteer maintainers of ex

More ›