matthewcford

Once you've done gov procurement, you can easily see why things like this cost this much. This is nothing.

I've heard of large SIs charging millions for discovery work, only for the report to say the budget is not enough to build the project.

Never mind the standards orgs tendering needs to meet (ISO 27001, Cyber Essentials Plus). It's not for the faint-hearted.

I've tried it, as even a low dose gives the person much more willpower to make other life changes, which have nothing to do with the effects of feeling fuller.

Ozempic and other GLP-1 agonists can:

Modulate dopamine release in brain regions associated with reward and motivation, such as the mesolimbic system and nucleus accumbens.

Reduce the rewarding effects of food and possibly other addictive substances, which may explain the increased sense of control or willpower.

Some studies in animals show that semaglutide can alter dopamine metabolism and suppress dopamine-driven behaviors.

I'm using nextdns - happy with it.

I am curious why proxying it through a managed domain incurs additional costs.

We've been building same thing hope to share it soon.

This is interesting; we've also been building an incident management tool, hoping to open-source it shortly, once it is ready.

The majority of this is underdevelopment; at the moment it does rate limiting and PII detection.

I used to live above the Wetherspoons, in the building next-door to the Angle; I don't miss the dunks pissing on my doorstep at closing time.

I do miss how central and convenient it was, though.

Likely to be a multi-threading issue; my bet is the cache client wasn't thread-safe. I've seen this in some apps before and the solution was to turn off multi-threading while we debug the library that was causing the issue.

Fresh is quite new imo it's not quite ready for a large project

M3

Can you run your app on a smaller vm but boot a larger one ok demand? I'm thinking bumblebee etc

I've seen issues with RDS after an update. We were seeing CPU max out; adding more readers didn't help, and in the end, we had to rebuild the entire cluster. This somehow fixed the issue, and CPU use returned to normal levels; we didn't make any code changes to the app.

Emphasizing that you have already done performance optimisation for ML algorithms would make me trust this a lot more, esp as you then open-sourced it.

The pricing was the same for self-hosted installs...

This is so that search can work, they index service start pages so they have results for services that are hosted on other domains. There is also something about having a consistance UX, at least to the point of handoff to the service (which might not have been updated yet). The start pages also provide the user with guidance were say the service might be only avalible at certain times, or if you need documentation ready before you use it.

can it sort by id?

Perhaps, I know around 5-6 agencies that have made this work.

I don't think they underestimate the effort or lack the skills; the issue is trying to do both - grow the agency and launch a product.

The ones which made it either split the agency or stopped doing client work.

I’ve run a 15-person agency for 14 years (bootstrapping), and the stress can’t be underestimated.

I know plenty of agencies who have either switched to products or just sold up / closed.

Using associates is a good option if you have a good network, however, it can be just as stressful if they don’t deliver.

If github are using private repos for the LLMs I think that would be a serious breach of trust.

We had a RDS db running at 100% when typical traffic is say 40/50%.

Support never found the solution, we had to scale up and out and completely replace the db two times for performance to return to the baseline.

All costed about 10k for what was essentially an aws bug.

Dont grow past 1 person, unless you love sales, the end game is your just in sales.

We bill by the hour (5min rounded) and only charge for focused work, if your rate is too low then increase it (if you can)

AWS Copilot https://aws.github.io/copilot-cli/

And it's a nightmare compared to fly.io

You update the toml file once (e.g if you have a release step), and then don't touch it again. Its more powerful than Heroku, so with that it gives you the option of adding additional config options.

I mean it even creates the basic file for you, I'm not sure what the issue is?

Will there be beta releases before Feb? I'm looking to kick off some headless shopify projects soon - and ideally I don't have to rewrite them in a few months time.

You know you can also use the heroku build packs on fly.io right?

N26 I see (as my phone died and needed to setup a new one) uses facial recognition to determine identity, you take a selfie video when signing up.

IMO this approach would be a good way to confirm identity over a sms.

What are the chances the microphone will be useable?

Looking forward to reactnative support.