Kamran Ayub

Extraction and decision models are exactly the kind of novel developments we wouldn't have thought of without first building on LLMs and running into their limitations. And there will be more deterministic-leaning models where these came from. blog.cloudflare.com/clairic-extr...

One wall section is ready! Sheathed and framed. Will do the other two next and tilt them all up in one session, with a helper... cause they are heavy 💪 I think I COULD do it myself but a wall falling on me is not my idea of fun. Better to be safe than sorry 😜

Before sheathing, which locks everything in place, I squared the panel by pinning the bottom plate to the subfloor flush with the rims (which I know are straight and square). Then putting four nails in each corner, same location, and measuring the diagonals. If they didn't match, I'd adjust the top.

To answer "what's happening right now" and "why is everything so broken" it's because it's way easier to sus out vulnerabilities with AI coding tools. See the Switch 2 userland hack. (Via Ars Technica)

Progress today: all the north wall window framing besides cripples above the header. That means tomorrow is sheathing day! And then I'll probably need some help tilting them up. 😅

Action shots!

Progress today: framed all the full height studs into each north wall section. Had one little issue where an edge got split but nothing some glue, a clamp, and a screw can't fix 😅 Next time those flat pieces get screws to avoid splitting. Still todo: cut and nail all the jacks, sills, and cripples

Today's progress: mostly prepping for the north wall framing, which is the most complicated 😜 Cut all the studs to length, cut sole/top/cap plates to length, marked everything for story sticks, and put up tilt stops. Also organized the wood into proper piles based on timing.

What, what! All the framing, sheathing, windows, and the door have arrived. When this is through, I will have an actual building! Pretty exciting. It doesn't seem real until this phase when walls go up.

Even if Ryan wanted to (and I believe he did earnestly want to), Deno could never and would never "beat" Node. Not if it had to be compatible. That's playing catch-up to follow the leader. That's not a winning strategy. It will be the same story for any fork that is pursuing that strategy.

There's still an opportunity for Node Foundation to do it themselves. I'd be interested in what they think about this.

Even if Ryan wanted to (and I believe he did earnestly want to), Deno could never and would never "beat" Node. Not if it had to be compatible. That's playing catch-up to follow the leader. That's not a winning strategy. It will be the same story for any fork that is pursuing that strategy.

Instead, in my opinion, a better strategy would have been to actually "merge" the best parts of Deno into Node core, basically making Node transform into Deno. It was already mostly compatible. If they wanted to, they could likely have approached Node Foundation with a plan to close the gaps.

Definitely will discuss this next week 🔥

DIY is never really about saving time, but it does save you money. I may be all-in for less than $30k compared to a turnkey installation of $70k or more (since even this Studio Home doesn't include proper roof insulation for Minnesota winters)

The cost is higher than what it quotes too because that assumes you pay it all in cash. Practically, you'd get financing and then you'll pay interest on top of the project cost over time, Studio Home quotes 7.25% at excellent credit with 10 year term. That's over $26k. So call it ~$100k all-in.

DIY is never really about saving time, but it does save you money. I may be all-in for less than $30k compared to a turnkey installation of $70k or more (since even this Studio Home doesn't include proper roof insulation for Minnesota winters)