now do crypto
766 karma · joined June 29, 2019
https://x.com/heysamtexas or
https://linkedin.com/in/heysamtexas/
The use of commit-hooks is also very clever (mentioned here in the replies)
Then those tokens would be saved in the database for your freelancer to use on their customer's behalf.
One left tech to join his brother in an outdoors adventure company where they rent/sell rafting and camping equipment and run tours.
I know another one who shutdown his zombie startup. Now he just manages his rental properties and runs a newsletter for a "ZeroHedge" type crowd.
No one seems to be getting rich but they seem much more fulfilled.
This is in my requirements.txt
git+https://github.com/richardpenman/whois.git@master#egg=python...
And the meat of the file:
import whois
"""Get WHOIS information for a domain."""
try:
domain_info = whois.whois(domain)
except whois.parser.PywhoisError as e:
return Response({"message": "Bad request", "details": str(e)}, status=400)
if domain_info["domain_name"] is None and domain_info["registrar"] is None:
return Response({"message": "Bad request", "details": f"{domain} not found or registered"}, status=400)
return {"result": domain_info}If this thing takes off and becomes a thing then adding those will be fairly straight forward.
Long story short is that I've decided to build an API service that wraps many of the most useful libraries and other services, even those "one-feature" SaaS offerings.
Be gentle, I'm sharing it to poke and critique.
I have some colleagues in F50 companies who asked to have a service like this deployed internally because of a number of bureaucratic reasons.
It is easier to get an internal API approved than it is to add it to the existing compiled/deployed software stacks. (weird, but it is what it is -- mostly about software supply chains)
Anyway, go look at the openapi spec and see that it is mostly just about scraping and format conversions (for now).
That didn't work for a number of reasons (cooking the books), but also network bandwidth is not fungible. Unlike commodities such as oil or natural gas, bandwidth’s value is highly dependent on specific factors like location, time, and network conditions. This variability makes it difficult to standardize bandwidth as a tradable commodity, complicating efforts to create a seamless trading market.
There are a few in the crypto/DePIN space poking at this problem. I remain highly skeptical.
I've implemented this in our workflows, albeit a bit more naive: when we kick off new processes the user is given the option to "put a human in the loop" -- at which point processing halts and a user/group is paged to review the content in flight, along with all the chains/calls.
The human can tweak the text if needed and the process continues.
It is a land grab like everywhere else.
The crypto people have already shown up, and the solopreneurs are chiming in as well.
Hopefully Bluesky’s algorithm can make the experience significantly different that people stick around.
Instagram launched a recent thing where they allowed people to reset their profile history and clean the slate.
I'm out of my depth here but genuinely interested and curious to see over the horizon.
David Tippet (formerly opensearch and now at Github)
A great podcast with David Tippet and Nicolay Gerold entitled:
"BM25 is the workhorse of search; vectors are its visionary cousin"
I dont know who that would be.
Multi-sided marketplaces are really hard to bootstrap, but with the right offering to the right people— it could be a win.
but seriously, this is a neat idea. and kudos to OP for building a full prototype -- this is the level of geekery I show up for.
I remember the days of T9 on the old Nokia phones. I was so good that I didn't have to look down at the screen.
My favorites were fixed in my phone. I knew how many down clicks on the button would land on the right friend/family member. I could literally send messages from my pocket.
And yes, I admit -- I did engage in texting while driving. And this is the part where I justify it -- "but I had eyes on the road the whole time!"
Oh! and there was old skits on late nite where people woudl compete with the old morse code guys.
Senior Python/LLM/Django Engineer & Technical Leader
"Let's do this" attitude + proven track record
└─── IMPACT
• 40K+ nodes/validators (petabyte-scale)
• $1M+ annual infrastructure savings
• 150B+ messages/month (200M+ MAUs)
└─── EXPERTISE
• Django, Python, LLMs, MLOps, Automation
• AWS/GCP/Azure, K8s, Terraform
• PostgreSQL, Vector DBs, Observability
└─── LEADERSHIP
• Technical architecture & strategy
• Team mentorship & process design
• ISO 27001, GDPR compliance
┌───────────────────────────────────────────────┐
│ Linkedin: https://linkedin.com/in/heysamtexas │
│ Github: https://github.com/heysamtexas │
└───────────────────────────────────────────────┘Having all that in place really changes my attitude from "playground on my machine" to "put my name on it and ship it." -- it makes me proud.
I'm heavy on python, so there are a few "aftermarket" bolt-ons:
- ruff for linting and formatting - flit for package management (will experiment with uv in the coming days) - commitizen for commit formatting - pyright for type checking (but not for Django projects) - pycharm is where all the magic happens - pre-commit for commit hooks and final checks before it get shared to the repo
I don't mention all the CI/CD stuff here because that is a separate thing.
Building up this stack took a while to onboard into. Taking it on all at once would be overwhelming and leave me bewildered.
It is nice to have claude or chatGPT around to have around for going into the weeds. However, "trust but verify" -- these LLMs will hallucinate, so you should check the original documentation.
The overall impact has been that I now use template repos where I can spin up a production-ready codebase with all the fixins. Not only that -- it is also ready for others to consume in a predicable fashion.
I still need to write docs, use cases, and things like that. But having this in place give me the opportunity to run faster, ship faster, and revisit old work with a confidence that was not there before.
In about the length of a weekend I was able to find a reasonable group of settings for my pyproject.toml and roll that out to many of my projects.
The tooling is maturing and converging on more common ground (eg pyproject.toml) while adopting existing, accepted practices and standards.
The documentation is awesome and integrates well within my IDE of choice (pycharm)
Myself, I run ruff with most of the lintiners turned on. It is quite strict and opinionated (to the point of pedantic) -- a good learning path for me because I would never read PEPs anyway.
But once it is running I just stay within bounds and take the enforced rigor around doc-strings, syntatic sugar, and other little things.
I'm not on the uv train yet because I have no issues with venv and pip right now.
1. https://github.com/simplecto/django-reference-implementation -- My personal production-ready Django boilerplate. "There are many like it, but this one is mine"
2. https://github.com/simplecto/sitemap_grabber -- A python library to recursively crawl every sitemap.xml for a website. Also handles robots.txt and other well-knowns.
3. https://github.com/heysamtexas/django-oauth2-capture -- A Django app to capture OAuth2 tokens for non-authentication purposes, enabling your application to act on behalf of users across external platforms like GitHub, LinkedIn, and X (Twitter)
I'm also taking popular and helpful software and wrapping them in RESTful apis as part of a larger api project I call the JOAT (Jack Of All Trades).
4. https://github.com/heysamtexas/REST-headless-browser -- Playwright headless browser wrapped in a FastAPI REST application, running inside a docker container
We ingest your data wherever you point our crawlers and then clean it for work working in RAGs or chained LLMs.
One library we like a lot is Trafilatura [1]. It does a great job of taking the full HTML page and returning the most semantically relevant parts.
It works well for LLM work as well as generating embeddings for vectors and downstream things.
There are some good grifts on LinkedIn, X, and Substack if you want derivative news and "knowledge."
The leap from playing with ChatGPT to building Bayesian filters in scikit-learn was like going from the 1-meter to 3-meter diving boards (keeping with the swimming theme). I'm not ready to climb to the 5 and 10-meter versions of AI/ML yet. :-)
Staff+ Engineer with deep experience in SRE, Blockchain, and DevOps. Specializing in building processes and systems that serve as flywheels, accelerating team productivity and scaling infrastructure efficiently.
Solutions: SRE, Blockchain Infrastructure, DevOps
Services: High-impact technical leadership, Infrastructure automation, Process optimization, Team enablement
Tech Stack: Django, Python, Rust, Golang, Terraform, Kubernetes, Docker
Cloud: AWS, GCP, Azure
Blockchain: Ethereum, Solana, Zk-rollups, Oracles
Specialties: CI/CD pipelines, Scalable infrastructure, Compliance (ISO, GDPR)
Available for fractional engagements or project-based work.
----
Book me directly: https://calendly.com/heysamtexas
LinkedIn: https://linkedin.com/in/heysamtexas
----