jonnycoder

I recently used AWS Textract and had good results. There are accuracy benchmarks out there, I wish I saved the links, but I recall Gemini 2.0 and Textract towards the top in terms of accuracy. I also read that an LLM could extrapolate/conjure up cropped text therefore my idea would be to combine traditional OcR with LLM to determine conflicts.

I've taken part 1 of 3 in Andrew Ng's machine learning specialization which covers the math for supervised learning, linear regression, etc. As I started part 2 (neural networks), it built off the math from part 1 such as the sigmoid activation function. This is what I think of when the OP refers to learning ML from first principles. I highly recommend Andrew Ng's course and I feel like I need to take it again to really understand those basic building blocks.

I am doing OCR on hundreds of PDFs using AWS Textract. It requires me to convert each page of the pdf to an image and then analyze the image and it works good for converting to markdown format (which requires custom code). I want to try using some vision models and compare how they do, for example Phi-3.5-vision-instruct.

He just revoked Biden’s Eo that strengthened h1b

Notes are extremely helpful, including those about coworkers you meet. Write down every detail during or after your meeting. Use a mind map program (miro works ok).

With the power of LLMs, notes will be even more helpful as we come up with more innovative ways to parse our daily lives.

I have the opposite problem. I know exactly what I would do if I had enough money in the bank to live off the interest. I remind myself how lucky I am and to be grateful, which is probably the most helpful thing I do.

As someone who has developed software for almost 20 years, and who is also staying on top of LLMs because I genuinely enjoy it, I really feel like I will make some good money soon.

As a senior software engineer who was unemployed for over a year, I can confirm almost nobody is hiring USA Javascript and Python devs with 10+ years of experience with some big accomplishments. I got lucky with a backfill.

He says he has a hard time beating his feather in arm wrestling despite him working out. Anecdotally blue collar people have much strong wrist flexion (cupping) than us white collar people, but pronation and technique can help negate that. My experience shows that power cleans can help with arm wrestling but not many people do those.

I say convert it all to dispersed camping. Leave the zoos, I mean national parks, to the common overcrowding. Let us have our forests back. It’s obvious that their policies have led to many of it burning down in mega fires that get so hot it kills every living thing in its path.

This would be a fantastic chrome extension because Amazon would never do this. It would be great to vote on the reasons why to boycott, allowing the most egregious reasons at the very top.

I'm on the fence regarding a "likely boycott" for Ooni pizza ovens, specifically the Karu 16 dual fuel. There are many videos about defective or improperly installed thermocouples. Ooni has some really helpful FAQ guides for fixing it on your own, but I was amazed at how many videos exist about this problem for an $800 pizza oven.

It's just their new brand name. I get the impression they are bringing genAI to boomers in a user friendly way. While everyone of us on HN have used chatGPT and knows the pros/cons, the regular users need a more gentle introduction using the basic use cases such as proofreading and simple image editing.

I was underwhelmed reading the link about the upcoming Intelligence features, but I put myself into the shoes of my dad and thought... well I guess that's a good starting point.

I literally just stumbled upon the designer and he posted a tutorial on the fiery pricing bento boxes: https://x.com/alex_barashkov/status/1841493887359066421

I am blown away at all of his work. He has some graphics and animation that I want to mimic for a "solar" theme I am working on.

Location: McCall, Idaho

Remote: Yes

Willing to relocate: No

Technologies: Python, C++, C#, JavaScript, TypeScript, Django, Node.js, CSS, Tailwind, SvelteKit

Résumé/CV: https://jon-olson.com/Resume_Jonathan_Olson.pdf

Email: jonnycoder@gmail.com

LinkedIn: https://www.linkedin.com/in/jwolson1/

Hi, I'm Jonathan. I am a Principal/Senior Software Engineer with high impact projects in backend, embedded and UI.

I have worked as a team lead in a small startup-like environment at VIZIO building embedded features in Python, Django and C++ and deploying my code to tens of millions of consumer televisions. I have also worked and lead teams at Intel gaining experience with a high degree of professionalism, communication, and quality.

I embrace any and all technical areas because I love solving problems. For example, without any experience in embedded development, I was offered to write the bluetooth pairing for VIZIO's first voice remote. I love building products and I love gaining experience at lower levels of the stack.

Pay rate is not a concern but it is important to me to be a helpful coworker and to also work with those who are positive and enjoy what they do. Our industry has been gaining a bad reputation but I know there are fantastic and friendly people out there - I've worked with some of them, and I strive to role model their behavior.

How does this intersect with evaluation in LLM integration & testing?

If you mean by mass producing energy dense processed foods that are much cheaper than meat and fruit, then yes I agree

Im on 16 months, definitely be careful

Seems like organized chatGPT in the form of mind mapping. It’s quite intuitive to me because I’ve had some chats where I kept scrolling back to the first gpt response. Therefore, you can map out a question and answer, then create nodes for follow up about specific details. Each branch of the tree structure can organize a rabbit hole of follow ups on a specific topic.

SF is far from an “amazing” city. From Gdansk to Bangalore to NYC to Amsterdam, and many others, I put SF at the bottom of my list. And I love cooler weather.

Funny experiment, I asked ChatGPT 4 and I always responded with “that’s not correct” 2-3 times and it eventually answered with Seer Gilligan.

I have modal, runpod and beam.cloud in my notes.

Surprised not seeing modal recommended more. I plan on trying it it myself soon.

The MTEB leaderboard has you covered. That is a goto for finding the leading embedding models and I believe many of them can run locally.

https://huggingface.co/spaces/mteb/leaderboard

KISS doesn't get me employed though. I narrowly missed being the chosen candidate for a State job which called for Apache Spark experience. I missed two questions relating to Spark and "what is a parquet file?" but otherwise did great on the remaining behavioral questions (the hiring manager gave me feedback after requesting it). Too bad they did not have a question about processing data using command lines tools.

As the op, you can do both semantic search (embedding) and keyword search. Some RAG techniques call out using both for better results. Nice product by the way!

Yes very easy with a ground mount with 4 panels, a battery and inverter.

I should write an article about it because I had planned to build something myself and keep all equipment outside and build a small wooden enclosure away from the house for it.

While the article probably doesn’t use UL listed equipment, I’ve found that equipment exists that does such as eg4. They run 48v which makes it legal in addition to being Ul listed. It’s typically the installation between solar panels and power inverter/converter that needs to be up to code.

Check out the EG4 specs from signature solar. They have been reviewed a lot by Will Prowse on youtube and it’s what I would buy.

This seems similar to building a RAG router (1) to perform dynamic retrieval/querying over data.

After getting hundreds of questions on my Interactive Resume AI chatbot (2), I've found the user queries can be categorized as: greeting, professional skills question, professional experience question, personal/hobby question and common interview question.

I am currently working on building a RAG router to help improve the quality of Q&A responses. I currently use gpt3.5 turbo without any special RAG techniques and the quality is lacking on performing Q&A over my resume and Q&A csv file. GPT4 works well but is too expensive.

1. https://docs.llamaindex.ai/en/stable/examples/low_level/rout... 2. https://jon-olson.com/resume_ai

This is fantastic, I just subscribed!

I had a similar idea of scraping lever, greenhouse and linkedin to get informed of absolute latest senior software engineer jobs. I also wanted to correlate to past job posts to rule out duplicates/reposts, and to analyze against what I am looking for. Some jobs rule out certain states and timezones. Other jobs are primarily java which is my only hard-no.

That’s well said. I mostly use LLMs to help me learn faster, not to do my work for me. There’s a lot more value to applied learning but it does require a lot of time and effort.

I just learned React/Nextjs for a take home coding interview assignment. I used useState a few times. Why would I need redux when there’s useState and useContext?

Other than being overly verbose, I think React is ok after building my website with Sveltekit. There just seems to be way more support for off the shelf react components and 99% of jobs ask for React.