At the moment I have been pretty satisfied with the LLMs that Kagi is offering in their standard their for my low requirement queries. Supposedly most of the apis they use are too alternative servers that host instances of the models, some claiming they don’t use your data. I am not aware to what extend open source is serviced in those models, but I have little interest in Lumo for the time being. Can someone explain to me why open source matters for LLMs?

How do you expect anything to be completely free?

I’m using the AdGuard extension. Would this potentially be a preferred alternative?

As of brother Lance has been hearing my prayers. He just put a video out, really tripping my understanding of dialling in. Honestly, I have to admit I have been fixated about achieving high pressure, and it’s true the window of balance seems to be really small, resulting more often in mudy bitter harshness.

Going to give turbo a shot.

I put a temperature probe in the E61 grouphead to check the temp on the Mara X, and now i can coffidanlty say it does have stable temperature control. It reads a 1-2 Celsius higher in the E61 injection point, right about the shower, than what the manual says the temp ought to be in the various modes, so there might be some heat lose when the water hits the coffee. Several times I tried Mode 0 to potentially offset the high reading temperature, it doesn’t seem to be a good match for the medium roasts.

with these precision baskets, targeting a shot time first seems to go out the window. Like you also say, go off taste is what i try. But then changing variables don’t always result in my intended effect.

I like to believe i have my puck prep pretty much dailed in, at the very least consistent.

An obvious variable many like to point out are good quality fresh beans. I always buy from reputable local roasters. Buy in batch, rest for 1-2 weeks, freeze, and take out a week’s worth of beans as i go. There might be a bad batch occasionally, but i seem to screw up on the simple blend roasts too.

the crema is indeed non-existent on these beans, I have seen the same on other beans I used that are on the older side. I am not a crema hunter, sure looks nice, but when there is plenty I dilute it with the rest of the cup usually. Of course in this case, might be an indicator that the beans are a bit stale.

You give a good suggestion, I might indeed give the beans a shot in my Clever dripper instead, even though there are ‘espresso roast’. They are definitely very lightly roasted for espresso. I might be chasing a brew method that doesnt yield the best result anymore.

These DAK beans aside, all my other frustrations are still there. I had luck this morning, grinder a bit finer brewing a blend for my cappuccino. And I have a ton 3x1 kg of brazilian roasts, from classic medium roast to anaerobic fermented, to try out the spectrum.

I am getting a puck screen in the mail soon, in order to eliminate some of the headspace with lower doses. In the 17-19 grams pullman basket, i consistently seem to get better results with 19-21 gram doses, but i don’t want to drink and waste so much coffee.

How is the heat production and ticking noise on the Belkin? I have been pretty content with a qi1 anker magnetic puck charger, but I read the qi2 from them gets pretty hot.

How quickly those this flood your storage?

Currently I ran invidious in docker. The solution is … just okay, performance isn’t great. But at least you only stream whatever you pick for watching.

I only follow around 15 channels. For simplicity 10 video are added daily to my subscription feed.

Are there privacy concerns with this solution. This together with ad blocking were my main drivers for invidious.