dimitri-vs

6 ms·

61.
by dimitri-vs 8mo ago
This assumes an educated, passionate and patient user that 99% of people are not. They wont ask for a hammer - they will ask for a rock tied to a stick and get pissed off when it doesn't work like a hammer. They will ask for paint that
62.
by dimitri-vs 8mo ago
The turn-key option is ostris ai-toolkit which has good tutorials on YT and can be run completely locally or via RunPod. Claude Code can set everything up for you (speaking from experience) and can even SSH into RunPod.
63.
by dimitri-vs 9mo ago
I wouldn't say the US gov has a good track record regarding it's "job" in healthcare.
64.
by dimitri-vs 9mo ago
In theory, yes. In practice the shit data you are working with (descriptions that are one or two words or the same word with ref id) really benefit from a) an agent that understands who you are and are likely spending money on b) has access
65.
by dimitri-vs 9mo ago
TBH I'm not building "production grade" apps depended on by hundreds of thousands of users - our clients want to get to a live MVP as fast as possible and love the ability to iterate quickly. That said, it's well know th
66.
by dimitri-vs 9mo ago
Reality is we went from LLMs as chatbots editing a couple files per request with decent results. To running multiple coding agents in parallel to implement major features based on a spec document and some clarifying questions - in a year. E
67.
by dimitri-vs 9mo ago
I'm in Claude Code 30+ hr/wk and always have a at least three tabs of CC agents open in my terminal. Agree with the other comments: pretty much running vanilla everything and only the Playwright MCP (IMO way better than the native
68.
by dimitri-vs 9mo ago
This is me. Was a huge Cursor fan, tried Claude Code, didn't get it, tried it again a year ago and it finally clicked a week later I cancelled my Cursor sub and now using VS Code. I don't even like using CLI, in fact I hate it, bu
69.
by dimitri-vs 9mo ago
Chandra
70.
by dimitri-vs 9mo ago
Yes, all the major CLIs (Claude Code, Codex, etc) and many agentic applications use a large model main agent with task delegation to small model sub-agent. For example in CC using Opus4.5 it will delegate an Explore task to a Haiku/Son
71.
by dimitri-vs 10mo ago
As someone that's currently building accounting (and many many other) tools for myself: yes, it can. But with a big fat asterisk that you: 1. Need to make it aware of all relevant business logic 2. Give it all necessary tools to iterat
72.
by dimitri-vs 10mo ago
FWIW most LLMs are pretty terrible at estimating complexity. If you've used Claude Code for any length of time you might be familiar with it's plan "timelines" which always span many days but for medium size projects get
73.
by dimitri-vs 10mo ago
Correct me if I'm wrong but I think the new "skillss are exactly this, but better.
74.
by dimitri-vs 10mo ago
Agreed. Been using Claude Code daily for the past year and Codex as a fall back when Claude gets stuck. Codex has two problems: it Windows support sucks and it's way to "mission driven" vs the collaborative Claude. Gemini CLI
75.
by dimitri-vs 11mo ago
Agreed, but I'm sure you can see why people prefer the infinite patience and availability of ChatGPT vs having to wait weeks to see your doctor, see them for 15 minutes only to be referred to another specialist that's available we
76.
by dimitri-vs 11mo ago
I do dozens of transactions every month sending payments to various freelancers. Been doing this for five years and can count the numbers of times I hand problems making payments on one hand - all we're minor and resolved in just a few
77.
by dimitri-vs 11mo ago
Agreed, to the point where I built my own UI where I can simultaneously generate three images and see a before/after. Most often only one of three is what I actually wanted.
78.
by dimitri-vs 1y ago
In the case of AI, it gets even worse when you factor in MCPs - which, to continue your analogy, is like letting random people walk into the factory and adjust the machine parameters at will. But people won't care until a major correct
79.
by dimitri-vs 1y ago
I thought we had all collective moved past the naive idea that any corporation is ever "your friend".
80.
by dimitri-vs 1y ago
Every time I overreached with trusting AI to correctly implement large parts of code I've paid the price in lost productivity later. I imagine with fully vibecoded apps the skill you need most is "you know what maybe I didn't
81.
by dimitri-vs 1y ago
As someone that has used coding agents extensively for the past year, the problem is they "move fast and break things" a little too well. Turns out that the act of writing code makes you think through your requirements carefully a
82.
by dimitri-vs 1y ago
It's not you, it's the productivity influencers making you think it's "supposed to be" more than what it is: a nice UI to edit a collection of markdown files. I realized this when I opened my Vault in Cursor/VS
83.
by dimitri-vs 1y ago
IIRC the SWE bench dataset gives you the full repo snapshot + the issue text, the evaluation pipelines typically run some kind of retriever (eg. grep, BM25) to pick a subset of files to place in the model’s context. They provided context is
84.
by dimitri-vs 1y ago
This effectively kills this benchmark.
85.
by dimitri-vs 1y ago
The only acceptable response to obvious AI slop - unless it's it's clear it's been heavily reviewed and updated - is to put it back into the AI and ask it for a 1 paragraph summary and work off of that.
86.
by dimitri-vs 1y ago
Yes: https://github.com/blakeblackshear/frigate/releases Ctrl+F: "Face Recognition" > Turn on face recognition & upload your first face via Face Library → Add Face. > Train and improve accuracy
87.
by dimitri-vs 1y ago
Not in my experience. Unless you explicitly prompt and bias that model for that kind of deep answer (which you won't unless you are already experienced in the field) you're going to get some sycophantic superficial dribble that&#x
88.
by dimitri-vs 1y ago
I haven't found such a bug yet. If it fails to debug on its second attempt I usually switch to a different model or tell it to carpet bomb the code with console logs, write test scripts and do a web search, etc. The strength (and weakn
89.
by dimitri-vs 1y ago
I don't think so, look at how Sora changed every... Well Operator was a game changer for.. Hmm, but what about gpt-4.5 or PhD level o3... o3-pro...? I mean, the 10k/mon agents are definitely coming... any day now... Anyway, I&#x27
90.
by dimitri-vs 1y ago
Have you tried datalab-to/marker with the "Use LLM" option? They have a playground you can test it out on https://www.datalab.to/playground but I use their local CLI option: https://github.com&#x2F

More ›