The best one was the 2048 era: https://hn.algolia.com/?q=2048

Meanwhile I have seen a lot of local restaurants advertising how ordering through their own delivery/pickup service can be 30% less than Uber Eats or DoorDash on their takeout boxes (and consequently in your uber/doordash order). It’s interesting that some of the POS companies are getting into the market: https://www.toasttab.com/local

There’s huge room for disruption in this model when a $12 chipotle order costs the consumer $30 after higher menu prices + delivery fees + tip.

When I worked at a restaurant that did Ubereats/door dash/GrubHub, we had a 15% upcharge on order through those platforms.

We would tell repeat customers they could use our website/app/phone to order directly and save the 15%. Almost no one converted. Convenience really is king.

So what this POS company is doing that’s kind of interesting is they have a mobile app just like Uber Eats/Doordash but it seems to be direct through the restaurant: https://apps.apple.com/us/app/local-by-toast/id1362180579

Actually when I ordered through it I paid $45 instead of $60 via DoorDash/Uber Eats but the order was delivered by Uber.

The reason why I think this version will work is because it’s the same level of convenience in a centralized mobile app so single payment method, single portal to browse restaurants etc.

(I have no association with this company I just was looking into this last week on the train)

Win-win then, since you made the upcharge.

Anything you can do with 1 person these large companies will figure out how to do with 10+ people + a PM + a Director/VP sponsor + 3-6 months of red tape and review and a late game “re-alignment” (rewrite) of the features right before the launch date.

More money is spent at most of these companies coordinating work than actually doing work.

I’ve also thought about this moreso as part of a workflow for quickly prototyping stuff. At least for a lot of the modern JIT languages I believe their startup times will be dominated by your imports unless you go with a fastcgi model. This came up as I started adopting h2o web server for local scripts since it has clean and quick to write config files with mruby and fast-cgi handlers and is also crazy fast: https://h2o.examp1e.net/configure/fastcgi_directives.html

Another place this can be useful is for allowing customers to extend a local software with their own custom code. So instead of having to use say MCP to extend your AI tool they can just implement a certain request structure via CGI.

An MCP frontend to CGI programs would not be a bad idea for a end user environment.

This makes me wonder if an MCP service couldn't be also implemented as CGI: an MCP framework might expose its feature as a program that supports both execution modes. I have to dig into the specs.

Fastcgi kinda loses all the benefits of cgi though.

From what I gather it has an embedded http control plane, yaml/json config for plugins, prometheus integration, and distributed compaction workers on separate, potentially serverless, hosts.

Yes, you have exactly got the key points.

It’s RocksDB but faster because data can be searched while still compressed allowing you to load more records in less cache/ram leading to up to 10x performance of RocksDB. It adds an embedded http control plane as well as supporting other extensions like MyRocks (MySQL) and Todis (redis compatibility).

Or at least thats what I got from it correct me if I am wrong rockeet.

OK. So "persistent key-value store" makes sense. But the "for External Storage" bit doesn't? What other kinds of persistent key-value stores are there? Or is "for external storage" just a synonym for "persistent"?

I love this app for bird sounds. I found ChatGPT is pretty accurate for pictures of birds or plants as well.

I'll try that. I like Seek by iNaturalist for plants and mushrooms. Still looking for better performance on trees.

I had the same experience, my typing speed with two hands is 90-120 but with one hand i can still get 50-70. The hard part is punctuation but with AI these days you could try just prompting and let the AI deal with syntax for you.

Right, you could run your text through an LLM that adds punctuation and then fix up the result manually. Would probably save time + fatigue

That's a fast speed. I practice to get to 60 then said well, that's good enough!

I have some thoughts on this (in the context of modern SaaS companies).

The most expensive parts of fixing a bug are discovering/diagnosing/triaging the bug, cleaning up corrupted records, and customer communication. If you discover a bug in development or even better while you are coding the function or during a code review you get to bypass triaging, customer calls, escalations, RCAs, etc. At a SaaS company with enterprise customers each of those steps involves multiple meetings with your Support, Account Manager, Senior Engineer, Product Manager, Engineering Manager, Department Manager, sometimes Legal or a Security Engineer and then finally the actual coder. So of course if you can resolve an issue (at a modern SaaS company) during development it can be 10-100x less expensive just because of how much bureaucracy is involved in running a large scale enterprise SaaS company.

It also brings up the interesting side effect of companies adopting non-deterministic coding (AI Code) in that now bugs that could have been discovered during design/development by a human engineer while writing the code can now leak all the way into prod.

The bureaucracy involved is usually the biggest cost driver. Another is the refactoring needed once mode code has been built atop the buggy code.

I’ve been doing this as well it also works well when you hook it up to Edgar or feed in investor relations documents or earnings transcripts. You can extract a lot of data at scale for regressions using small models with few shot prompts running locally.

What do you mean by "regressions"?

Scraping the documents with llm to generate numerical metrics and applying statistics on numbers

OP likely means statistical regressions.