haute_cuisine

Setup 2-3 accounts to use codex. Claude requires at least max plan to avoid hitting limits constantly.

Snowden already showed what lawful use actually means.

It's interesting that you frame the question as what is wiser. Here is a quote from Naval about making decisions, I hope it would bring more clarity.

Decisions:

- If you can’t decide, the answer is no.

- If two equally difficult paths, choose the one more painful in the short term (pain avoidance is creating an illusion of equality).

- Choose the path that leaves you more equanimous in the long term.

Also, never outsource decision making to anyone who doesn't bear the consequences of the decision.

You still need to register and put your credit card.

Can someone explain why Dario is making a public statement about this? It's also interesting that they use abstract we / they without putting exact names.

Not all projects are hosted at github. You also might want to receve relevant mail from fellow developers.

Only free individual can have strong ethics. There are no free people in capitalism, money is debt after all. Think of applied pressure once you sign under VC money and amount of brainwashing / gaslighting. I sincerely hope my observation is wrong.

I think the main problem is tech workers finally became a commodity. It was a brief time when hiring was hard and it was possible to get some sense of power because you could freely walk away or because you were a craftsmen.

Over the years, power slowly returned back to management and the industry figured out how to slice the creative role into small cogs: QAs, FEs, BEs, SREß, OPSs, POs, PMs. It's ten people now instead of three crafsmen. They can now follow a very strict process to produce average software with a lot of overhead. Customers don't care because one half is forced to use it and the other don't have a taste of what good software is. Partially, because they use abysmal MS products. Apple briefly showed to an average Joe what is possible when people care, but they're slowly losing it.

Now it's node/js everywhere and fierce competition being a ticket taker. People with no passion who are here just for the money and who don't care about the code. People who are here to play promotion politic games. Coding is a regular job now which pays decent and has low barrier to entry. You can fake it very far now with AI assistance.

Personally, I find AI generated text disengaging. A also heard some professional writers immediately notice disorganized storytelling patterns. Have you found a way to fix this? Is there any soul in generated texts?

email works, not everything has to be run by corporate greed

Is landing page generated by AI?

The problem here is that companies artificially limit integrations, so it's impossible to exchange messages between different providers like how email works.

It's most likely already installed when you send an invite to someone and they already used to how it works. It just works most of the time, well, besides slightly buggy text editor and almost non-working calls.

Pangram says it's 100% AI generated.

At some point social networks had XMTP servers, so it was possible for users to chat across social networks as protocol supports federation. It was even possible to use iMessage as a client.

Sadly, XMTP died once platforms gained enough users. The protocol itself is quite complex because of custom extensions (XEP) and different clients supporting different set of extensions.

I wasn't able to find a good desktop/mobile experience where I looked into self-hosting the XMTP server.

This is amazing! Thanks for sharing. What would be the elo gain for 4KB engine?

P.S. I assume 1200 elo in chess com scale (not lichess / fide elo) and bullet chess variant?

I love TUIs and spent some time exploring the project. I guess you posted it here to hear some feedback.

The landing page compares it to xtermjs as superior version. What I find strange, for example, FrankenTUI has a code explorer tab, but it's impossible to select any text or interact with a scrollbar. The "hotspots" section in the codeview misinterprets mouse clicks, highlighting line that is different from a click.

If it meant to be keyboard-first, I find focus management implementation unintuitive and shortcuts hard to discover.

It'll be still broken as it'll be claude bugs + frankentui bugs.

I checked your website, it's quite impressive to release 1-3 opensource projects per day.

Any released projects yet?

I mean, if you look visually it looks like it's working. Once I tried interacting with it, everything is broken in a subtle way. Website has the same vibe, it has nav bar, it does it's thing, but it's broken in many different ways (media breakpoints, animations, etc).

I think companies who buy software by feature checklists would love this sort of products.

Doesn't seem to be the case, gpt 5.2 thinking replies: To get the car washed, the car has to be at the car wash — so unless you’re planning to push it like a shopping cart, you’ll need to drive it those 50 meters.

Amazing work! Congrats on the launch. I’ll definitely give it a try

My first thought that it was made for companies which tie "AI usage" to performance evaluation.

Learn B2B sales process (prospecting, lead generation, lead data enrichment, outreach tools, etc).

What about SSH? Does it work? If yes, you can use some TUI browser as it would only pass updated SSH screen

There were some popular SaaS products doing aggregation form multiple sources (email newsletters, websites, rss feeds) into a daily morning email. One had the word "brew" in the domain.

I believe founders closed their service and started a new one based on AI (some yers ago).

It wound be nice to have such tool that can sort through communication and highlight important action items, I want this, but I want this to go further: I define my goals and priorities and it acts as personal assistant with preparing weakly goals, daily action items, managing my todo lists, outsourcing some part of my todo list, etc.

The important part is I can't just hook up some random saas into all my business communication.

Thanks for the writeup on the libvirt setup. At some point I used local docker containers for this.

What about `docker run`? It'll be the same isolated container that keeps state. You can also mount some local directory

Wow, that's fast. Do you know if committer's email also goes into calculation of a git hash? If it is, it should be possible to manipulate git hash in a very discrete way through the email address like this: user+<nonce>@example.com