mrinterweb

A lot of money and years of marketing the cloud as the responsible business decision led us here. Now that the cloud providers have vendor lock-in, few will leave, and customers will continue to wildly overpay for cloud services.

Nothing much. How bout you?

This seems like a good opportunity for a browser company like Mozilla to offer a GDPR compliant library that is easy to integrate that automatically applies user privacy preferences instead of showing the GDPR prompt. Opensource the library, and promote it. Try to make it an open protocol so other browser can implement this.

To be real though I'm sure that many sites would not want this because they rely on GDPR fatigue and users to just accept instead of taking a few seconds to opt-out.

One thing that frustrates me with estimating performance on AWS is that I have to dramatically estimate down from the performance of my dev laptop (M2 MBP). I've noticed performance tradeoffs around 15x slower when deployed to AWS. I realize that's a anecdotal number, but this is a fairly consistent trend I've seen at different companies running on cloud hosting services. One of the biggest performance hits is latency between servers. If you're server is making hundreds or thousands of db queries per second, you're going to feel that pain on the cloud more even if your cloud db server CPU is relatively bored. It is network latency. I look at the costs of AWS and it is easy to spend >$100,000 month.

I have ran services on bare metal, and VPSs, and I always got far better performance than I can get from AWS or GCP for a small fraction of the cost. To me "cloud" means vendor lock-in, terrible performance, and wild costs.

Investing in AI infrastructure seems very risky to me. With how short of a lifecycle computing hardware generally has, does it make sense to blow billions on hardware that will need to be replaced in a few years? AI companies have been burning billions on hardware. Will the costs of that hardware be covered by profits before the hardware lifecycle is up? That's a big gamble.

We used to talk about AI as having no moat (easy for other players to accomplish similar AI achievements). China has made that clear. Some open weight LLMs are fine and can run on laptops. It seems the new moat is model parameter size and VRAM requirements. I would bet on innovation in hardware or disruptive algorithms changing that game so better LLMs can run on more personal computers. Remember how bitcoin miners all used GPUs then ASICs came along and made that no longer profitable?

There are many ways the AI industry can be disrupted which makes it that much more volatile.

I can't tell if you're being sarcastic. Read either way, it works :)

One training source for LLMs is opensource repos. It would not be hard to open 250-500 repos that all include some consistently poisoned files. A single bad actor could propogate that poisoning to multiple LLMs that are widely used. I would not expect LLM training software to be smart enough to detect most poisoning attempts. It seems this could be catastrophic for LLMs. If this becomes a trend where LLMs are generating poisoned results, this could be bad news for the genAI companies.

Shocking no one, Google pulls the app too. https://www.404media.co/google-calls-ice-agents-a-vulnerable...

A government probably wouldn't have to push very hard for Google to revoke a dev's signing key, blocking apps signed with that key. So I don't see a difference. Technically sideloading will be possible, but Google will still have control of what apps can run on devices.

Google is making it so sideloading APKs on Android is no longer possible. This is one of many great reasons why that should not happen.

I want to like Ghosty, but not being able to search has been a deal breaker. Guess the alternatives to search are use tmux or grep the ghosty log, but tmux and I have never clicked, and searching the log is not user friendly for how often I find myself wanting to search terminal history. Ghosty has a lot of features I would want, but somehow it is missing search. https://www.reddit.com/r/Ghostty/comments/1i0g2ge/how_to_fin...

Maybe this concern doesn't apply to libghostty, as I think search would be more of a UI concern.

You and I should talk :) I've been thinking of this ever since seeing the movie "The Incredible Burt Wonderstone". He uses a sleeping gas on his audience packs them into a moving truck and "magically" transports his audience to a new location. Tada! Basically same idea, but substitute moving truck with jet.

How would a heads up email look like a phishing email? Blindsiding the maintainers like this is just cruel.

I would say react being the default expands to apps that normally would work perfectly server-side rendered. The insane amount of added boiler plate associated with writing an API, tests for the API (including contract tests), API documentation, API versioning concerns, deployment timing considerations; front-end API integration, front-end state management, front-end tests, API mocks, I feel like there's about 10 more items I could rattle off.

I feel like people forget that web apps can be rendered server-side, and with HTML-over-the-wire (HMTX, Rails Hotwire, Phoenix LiveView, Larvel LiveWire, etc), server-side rendered apps can have a UX similar to a react app, but with far less total effort.

I've played a couple hours of silksong. I don't get the hype. Its a fine game, but I really think people are over hyping it. The internet hype loop on this game is turning me off on it. It's a nice metroidvania game.

DisplayPort is the interface most gaming monitors use. If you don't need ARC or CEC, mostly used for home theater builds, DisplayPort is preferable.

I don't think I realized how many alternative formats there are for both 3339 and 8601. Most of those alternative formats are redundant. I like that this chart shows the venn diagram of where these standards intersect.

I'm certainly not advocating that "conspiracy to impede" should be illegal. I just feel that putting "impede" and "injure" together into a single charge seems like a way to deal excessive punishments in a way that could be very broadly interpreted.

Acceptance is embracing a fascist autocracy? "Hmmm. Libertarianism didn't resonate with liberals. Guess we will just go full on fascist.", is how I interpreted that.

"Conspiracy to Impede or Injure Officers" seems like a pretty broad category, and should be separate things. Conspiracy to impede is one thing. That might be applied to planning to be at a protest where you may be in the way of an officer. Injuring is a very different thing. Conflating the two into a single charge seems disproportionate. This conflation seems intentional to give an excuse to just round up anyone who plans to be at a protest, and charge them with a crime that has a punishment equal to assaulting an officer.

It is infrequent that "find usages" and "go to definition" don't work for me on good sized ruby code bases. Both solargraph and ruby-lsp seem to work fine for me. Occasionally, I'm surprised when they don't work. In my recent mileage, that happens once every couple weeks, and I use find references and go to definition multiple time per-hour.

Metaprogramming has fallen out of fashion in the ruby community for the most part. Rails is a great example of where that shift has happened. ActiveRecord used to have `find_by_<attr-name>(<value>)`, which became `find_by(<attr-name>, <value>)`. Sure devs can still go wild with metaprogramming in ruby, but it is generally discouraged.

(CMD|CTRL)+TAB is all I usually all I need. I know this may sound wasteful considering that I have two eyes, but I have only ever been able to focus my eyes on a single thing at a time. I usually make the window take up the full window, and I just switch between windows. It is fast to do this. I tried multi-monitors, but it wasn't for me.

Hopefully, OpenAI makes their APIs more affordable. So far, there are alternative LLMs and services that both outperform and are a fraction of OpenAI's pricing. OpenAI is usually one of (if not) the most expensive option, maybe that's because of the brand identification. Not really sure why people pay that premium.

Is a perplexity visit not cached and shared between users who perform a similar search? I don't know much about perplexity, but I'd be surprised if scraped results weren't used to serve multiple searches and users. By bypassing the no-crawl directive, that is a violation of the website's expressed request. I think it is different if individual users chooses to bypass certain things on a website, but for a company to choose to do it is another story.

I don't know about that. I've had my RTX 4090 for nearly 3 years now. If I had a script that provisioned and deprovisioned a rented 4090 at $0.70/hr for an 8 hour work day for 20 work days per month. Assuming I get 2 paid weeks off per year + normal holidays over 3 years.

0.7 * 8 * ((20 * 12) - 8 - 14) * 3 = $3662

I bought my RTX 4090 for about $2200. I also had the pleasure of being able to use it for gaming when I wasn't working. To be fair, the VRAM requirements for local models keeps climbing and my 4090 isn't able to run many of the latest LLMs. Also, I omitted cost of electricity for my local LLM server cost. I have not been measuring total watts consumed by just that machine.

One nice thing about renting is that it give you flexibility in terms of what you want to try.

If you're really looking for the best deals look at 3rd party hosts serving open models for the API-based pricing, or honestly a Claude subscription can easily be worth it if you use LLMs a fair bit.

I often provide mermaid diagrams in my promt. Mermaid seems to be a good common markup to communicate relationships between humans and LLMs.

Knowing that your content is unlikely to be seen by many humans, and is mainly just feeding an AI is pretty discouraging. AI is great at detecting patterns, if content creations is discouraged, and more AI content is generated, AI content is likely to become so homogeneous that the internet becomes more of an echo chamber (than it already is).

Scaling concerns can be a legitimate reason for a microservice, but I think those scaling concerns should be proven and not assumed before a new microservice is born.

I also hate the agreement of maybe one day we might... as a justification for a new microservice. The number of times that premature optimization didn't pay off is far less than I've seen it come to be.

Microservice should be an exception, not the preferred architectural design pattern.

Sometimes I cynically think system architects like them because they make their diagrams look more important.

For your example, rails apps handle this case by default with job queues that are managed by the deployment separately. There is a way to force the job queue to process in the same process as your web server, but that's not the way most should be running prod rails apps. There usually isn't anxiety associated with rails app deploys, in my experience.

One of the reasons microservice architecture originally became popular was to break apart monolithic applications. In many cases, I bet a big driver was a lack of separation of concerns, and a more modular design was desired. There are many ways to put up walls in software to help make software more modular and self-contained. Rails engines are a good way to make more a rails app more modular. The number of times I've seen microservices created for the purpose of modularity (not scaling concerns), and the complexity that has brought has really soured me on microservices.