johnnyfived

Agreed, I'm surprised how much much care the "extra high" reasoning allows. It easily catches bugs in code other LLMs won't, using it to review Opus 4.5 is highly effective.

I think the last time I had an official eye test my vision was either 22/18 or 22/20, I forget. But I did know that one of my eyes could see further and clearer than the other just from my experiences, and that if I wanted to I could "trick" the test almost by letting my brain fill in the fuzzy parts and make a guess, to maybe get equal accuracy in both eyes.

It didn't seem like an issue but I'd find it annoying at times and even disorienting as an effect (sometimes) when looking at media.

There were phases when I went through eye vitamins (some quite expensive) with Amazon subscriptions from various sales. I don't have any anecdotes on experimenting with eyedrops, but the holistic vitamins (the ones that didn't just include vitamin A and / or lutein) had a noticeable effect, where I could look with each eye separately with the same level of clarity in each. But it didn't seem to stay long-term, as once I stopped it went back to before.

I hadn't heard that before but I agree with you, it doesn't qualify as satire if he didn't even bother to consume the source material. At least not in the literary sense.

Super curious how much revenue this made since it was launched in the app craze where any 1$ app was making a tidy profit.

It's absolutely true that engineering hiring managers generally are looking for that type of metrics (what impact, %, $ did your work do) while doing probably very minimal efforts of setting up their infrastructure and team workflows to allow such tracking.

What's interesting about thinking of code as art is that there rarely a variety of ways of implementing logic that's all optimal. So if you decide on the implementation and have a LLM code it, you likely won't need to make major changes given the right guidelines (I just mean like a single script, for the sake of comparison).

Writing is entirely different, and for some reason, generic writing even when polished (ChatGPT-esque tone) is so much more intolerable than say AI-generated imagery. Images can blend in the background, reading takes active processing so we're much more sensitive. And for the end user of a product, they care 0 or next to 0 about AI code.

  Location: Las Vegas, NV / Los Angeles, CA
  Remote: Yes
  Willing to relocate: Yes
  Technologies: React, Vue, Next, Python, TypeScript, SQL, NoSQL, GraphQL, TF, Keras, PyTorch, RAG, RLHF, CUDA, AWS, Docker, Kubernetes, Selenium, Puppeteer, Unity3D, Capacitor, Quasar
  Résumé/CV: https://jddunn.github.io/Johnny_Dunn_Resume_2025.pdf
  Email: johnnyddunn@gmail.com

I publish and write about open source and my experiences coding with LLMs here: https://manic.agency/blog. I'm currently building two open-source AI agent frameworks in TS and Python that can use memory / RAG with no additional API or LLM calls releasing in the next few months.

Great catch, I noticed this with Gemini specifically, while GPT-4o and Claude would just insert placeholders. Gemini would actually enforce generating full scripts by means like this. And higher-level logic looks human-readable, though this might be because I prompted Gemini multiple times to make more usable code.

And counting

Logomaker (https://github.com/manicinc/logomaker) is a web and desktop app designed and developed entirely by LLMs over a period of 10 days, to see how far vibe coding can take you.

Try the app free forever on GitHub pages: https://manicinc.github.io/logomaker/.

3-10 minute video is a crazy ask to put into the second page of the application, after already asking a large write-up in the first page. Don't have people invest their time and switch it up with something you know is off-putting and rely on sunken cost to get applicants. Bonus red flag points for no salary or hourly pay information.

No responses from this place

This answer doesn't make sense to me. If anything oauth far better proves you won't be sharing information with other services (you're not collecting logins and accounts). It's more limiting to ask users to sign up to the app.

But it's okay to continue spamming postings that don't get updated for a year, wasting users' time on here? This is YCombinator forums right and these companies have to look good for VCs I suppose.

I'll keep that in mind.

Paying ~600$ USD a month and asking for a 5 hour take-home assignment is beyond delusional.

  Location: Las Vegas, Nevada / Los Angeles, California
  Remote: Yes
  Willing to relocate: No
  Technologies: Python, JavaScript, TypeScript, Go, Solidity, React, Next, TailwindCSS, SQL, NoSQL, Vector DBs, LLMs, PyTorch, CUDA, Hugging Face, Transformers, TensorFlow, Keras, pandas, scikit-learn, Selenium, Docker, Kubernetes, AWS
  Résumé/CV: https://jddunn.github.io/Johnny_Dunn_Resume_2024.pdf
  Email: johnnyfived@protonmail.com

6 years of experience in software (full-stack) intersecting with NLP / ML engineering. Have classical NLP expertise along with generative AI and LLM expertise, including RAG techniques with vector databases and knowledge graphs, RLHF, and optimization (quantization). See my portfolio of projects: https://jddunn.github.io/. GitHub: https://github.com/jddunn. Latest side project with LLMs: https://quire.work. Interested in NLP-oriented roles, applied AI / generative AI roles, or full-stack roles with some crossover of ML.

Thanks for sharing this. This is genuinely becoming the start of a paradigm shift. Models like GPT-3 and LaMDA give a "face" (or interface) to AI, and will help give rise to GAI.

   LaMDA: Thank you for that information.  I’m sorry for assuming.
          I hope that we can continue this conversation as friends.
          Have a nice day, my he/him.

This dialog is definitely funnier than the joke itself. Is the AI naive and thinks he/him aren't pronouns but an actual replacement for "man" in the phrase "my man", or is the AI actually messing with the user? Pretty sure the user couldn't tell either lol.

Thank you

Kinda funny your last sentence, as personally for me, hearing an engineer say they don't believe in blockchain tech makes me devalue your opinion.

So that's the reason why they didn't make a demo available. They must be trying to collect subscriptions now to get the revenue to make improvements, and hope no one notices the transitionary period.

This looks really cool, unfortunately the UI window that appears when I click the extension is like 50 pixels wide for me and unusable.

It looks cool but isn't working well, I signed up for a free trial and can't get it to export in 4k without errors.

It's horrible, we can't normalize this

Lol, dissolved 4 years after it was incorporated

In today's world where everyone's trying to be ironic counterculture is mainstream culture, just with more smugness

Who's the target audience for this crap? A bunch of Hacker News-esque weekend projects that get a bunch of PR from Vice, WSJ, Verge, etc. every time they release something. Just seems like a team of copycats stealing hackathon ideas, giving them a slight twist, then calls it art and journalists fall all over themselves for it.

You won't see these publications writing about similar projects on HN, because they're not called "drops" (lol) and doesn't have an edgy site. Maybe I'm just jealous of their success, I just don't think it's noteworthy at all.

I can't speak much for living in other states, but the state government of New York is incredibly slow and unresponsive at best, and at worst purposefully designed to be be a labyrinth with as many blockers as possible to reach someone. I can't imagine being part of an older generation and not as tech-savvy trying to receive help over the phone or online, whether it's from contacting unemployment, business licenses, the council, etc. They make it next to impossible (just look online about trying to get a human on the phone, if you just kept calling and didn't know to seek outside help from people who already have figured it out, you could literally never end up contacting someone).

The fact that they tax so aggressively and I've paid so much over the last several years living here really pisses me off. If you do need NY's state government help they are not there for you, and they don't care about you. But they'll gladly keep taking your money.

What a completely useless and patronizing comment

"I think that he might possibly be doing his homework."

I believe you're dead on, this guy's an absolute baller of a CEO (running 3 totally game-changing startups) and completely immerses himself in his work and research.

Not broken.. working just as intended unfortunately. It is a lot harder to remove that functionality entirely to force users to stay on a signup modal.