HNHacker News
TopNewBestAskShowJobs

dalemhurley

643 karma · joined October 6, 2023

submissionscomments
dalemhurley··on Show HN: Waiting for LLMs Suck – Give your user a game
Thank you, added :)
dalemhurley··on Show HN: Waiting for LLMs Suck – Give your user a game
Maybe I should add slots to take it to the next level?
dalemhurley··on Show HN: Waiting for LLMs Suck – Give your user a game
Let’s keep adding more?
dalemhurley··on Show HN: Waiting for LLMs Suck – Give your user a game
Chase the high
dalemhurley··on Show HN: Waiting for LLMs Suck – Give your user a game
Agh damn, now it is released to the public it is too late.
dalemhurley··on An AI agent deleted our production database. The agent's confession is below
This post reads like “I gave the intern prod access and it is their fault”.
dalemhurley··on An AI agent deleted our production database. The agent's confession is below
Am I reading this correctly? You gave an LLM prod access? You told it that it was a staging env? The API key had the permissions to delete? You expect an API to have a confirmation message?
dalemhurley··on DeepSeek v4
Theoretically yes. It is entirely possible to poison the training data for a supply chain attack against vibe coders. The trick would be to make it extremely specific for a high value target so it is not picked up by a wide range of people. You could also target a specific open source project that is used by another widely used product.

However there is so many factors involved beyond your control that it would not be a viable option compared to other possible security attacks.

dalemhurley··on Claude Opus 4.7
Wild how many people were predicting the AI slop, but was dismissing it as unlikely beyond some trolls.
dalemhurley··on Is anyone actually using OpenClaw?
I think the issue for our community is we will just create a LLM wrapper app in a few minutes to solve what we need making OpenClaw have 0 utility.
dalemhurley··on Cursor 3
Thanks, it is horrible. This is a massive step backwards. The IDE provides so much extra abilities that an agent simply can’t handle.
dalemhurley··on TurboQuant KV Compression and SSD Expert Streaming for M5 Pro and IOS
Is t the point of an MVP to be an MVP?

The OP put together a POC and shared it, showing novel concepts used together. They are not some large R&D lab.

The purist tests being asked for is in contradiction to the ShowHN guidelines.

dalemhurley··on If you don't opt out by Apr 24 GitHub will train on your private repos
At least they are giving you the option to opt out, many other providers just trained on the source code.
dalemhurley··on Speed at the cost of quality: Study of use of Cursor AI in open source projects (2025)
I think the issue is people AI assisted code, test then commit.

Traditional software dev would be build, test, refactor, commit.

Even the Clean Coder recommends starting with messy code then tidying it up.

We just need to apply traditional methods to AI assisted coding.

dalemhurley··on Returning to Rails in 2026
Laravel is extremely stable and consistent.
dalemhurley··on Returning to Rails in 2026
The difference between JS frameworks and RoR/Laravel is the ecosystem cohesion. RoR and Laravel ecosystems employ the RoR or Laravel way of doing things and everything works together very smoothly.

JS solutions are loosely coupled, lots of good reasons to do so, but comes at a major complexity cost.

dalemhurley··on Returning to Rails in 2026
Ruby on Rails and Laravel all the way. Solid, proven, stable and scalable.
dalemhurley··on Don't post generated/AI-edited comments. HN is for conversation between humans
While I understand the sentiment, it ignores many people have English as a second language, many people are dyslexic and have dysgraphia. AI is a great assistant. A good approach will be to encourage people to develop their thinking than use the AI tools.
dalemhurley··on Techno‑feudal elite are attempting to build a twenty‑first‑century fascist state
That analogy does not make sense. You are assuming political spectrum is a left to right gradient which you just move along, when in reality it is a constantly shifting multi dimensional spectrum that shifts on different issues. However if we apply today’s centerist to the end of the 1990’s they would be more a Bush conservative.
dalemhurley··on We Will Not Be Divided
They are after the models without post training guardrails.
dalemhurley··on We Will Not Be Divided
My guess is their HR is already monitoring it with instant termination processes in place.
dalemhurley··on We Will Not Be Divided
XAI has already announced they are 100% in

https://x.ai/news/us-gov-dept-of-war

dalemhurley··on New accounts on HN more likely to use em-dashes
Several factors: 1. Em dashes are common use in the Queens English

2. People with dyslexia and dysgraphia can more easily interact online

3. People who speak a primary language other than English can more easily interact online

The last 2 options mean people who previously would have been more reluctant to participate now have less of a barrier.

So while there may be AI generated content, we should just assume it is all negative.

dalemhurley··on GPT‑5.3‑Codex‑Spark
1000 tokens/sec for a highly specialised model is where we are going to see agents requiring.

Dedicated knowledge, fast output, rapid iteration.

I have been trying out SMOL models as coding models don't need to the full corpus of human history.

My most recent build was good but too small.

I am thinking of a model that is highly tuned to coding and agentic loops.

dalemhurley··on SkillsBench: Benchmarking how well agent skills work across diverse tasks
That is my approach. I don’t think the papers author has actually used skills.
dalemhurley··on SkillsBench: Benchmarking how well agent skills work across diverse tasks
After several failures then a success I have the agent create the skill, next run it is successful first run.
dalemhurley··on What your Bluetooth devices reveal
Ring: thank you for the idea, "Introducing Ring Face-Off, face masks covering faces during a break-in is no an issue for Ring, we will track the thieves until they reveal their face to our Ring network."
dalemhurley··on GPT‑5.3‑Codex‑Spark
Yet investors keep backing NVIDIA.
dalemhurley··on GPT‑5.3‑Codex‑Spark
This is a win for agents, speed and intelligence is crucial to the loop. If the time and token cost is small you can iterate many times to correct mistakes.

Got to wonder why Wall Street is dumping NVIDIA.

dalemhurley··on Show HN: LocalGPT – A local-first AI assistant in Rust with persistent memory
I’m am playing with Apple Foundation Models.
← PreviousPage 5 of 10Next →