While half my timeline is arguing about €26K Macs, apparently all of OpenRouter's DeepSeek traffic fits on a handful of racks... I really think cheap open models we can run at home matter way more than who has the biggest box. If you've never tried a local model, just pick a small one and run it this weekend. It's way less scary than it sounds. Happy to help if you get stuck.
0
0
2
190
It turns out that the fastest terminal is one that just skips reading what you typed. BRILLIANT!
@valvesxyzlook at claims like this in case I can improve Ghostty. I grabbed your source and immediately spotted the issue: your read loop overwrites a 64 KiB ring. Everything older is discarded unparsed per tick. You're faster because you're not actually terminal emulating.
It means the nt number. In a kitten benchmark run at the rate you're claiming, roughly 6.6 MB arrives per 60 Hz tick and 64 KiB of it gets parsed. You're read(2)-ing 100% of the data and parsing about 1% of it.
There are other minor issues (a small fixed scrollback in a preallocated ring buffer, no grapheme clustering, throwing away data like OSC 8 URIs, etc.), but the major one is simply that you're discarding ~99% of the input.
For plain text, e.g. your cat benchmark, the final screen ends up looking right, but the scrollback is missing 99% of the file and each retained chunk starts mid-line. As soon as there are escape sequences in the dropped spans, your emulator is fully out of sync and invalid: a cut inside a CSI prints the parameters as text, and any SGR, mode, title, or alt-screen change in the gap is just... gone.
FYI, this is why I don't like when people point at refterm as a terminal benchmark. refterm implements a TINY subset of what a terminal emulator must support, it's closer to a fast file reader, not a terminal emulator.
I also don't mean any shade by this. I like to look at technology objectively to try to learn something and improve my own craft and work. So I'm sharing what I found.
I modified your code very slightly to parse on every full buffer (rather than per tick). And the results are attached. Both use main commit in each repo at the time of this post.
Note the graph is in log scale, the discrepancy between performance for things like Kitty images is measured in GB/s.
0
0
0
26
So Microsoft is letting 25% of their employees go, after they already let a lot go last year. I suspect this is just the beginning. I had hoped that these large tech corporations would reallocate their resources to innovation instead of just using their newfound AI capabilities to cut costs. Imagine if all these big tech corps had the same mentality as the Omarchy crowd. Then we really could fix everything!
0
0
0
22
Grok @bot keeps impressing me. One simple thing I use it for, which has already given enormous value, is scanning repos and tracking vulnerabilities. Every morning I get an e-mail from Grok Bot telling me if any unsecure code has been checked in, but more importantly it tells me if any of the used technologies have known CVE's. Stuff I might check myself every two weeks or so if i remember. Now I get the warnings automatically. It also has a list of all the networking hardware we use at work, and the version numbers of the software running on it, and tells me about any CVE's on that specific hardware.... oh and finally if any CVE's with a high rating show up, it informs me immediately 24/7. Simple use case, HUGE value for me.
0
0
2
201
Am I the only one who is seeing ChatGPT and Claude default to computer use for things where they already have direct access? Like checking or writing e-mail. Suddenly it wants to do Computer Use which is so inefficient compared to just using the direct plugin connection.
0
0
1
51
Life as a head of IT at a venue is never dull. Heading of to work now. Today I have a misbehaving server I need to get fixed, some vendors to talk to, a call with Salesforce, discussions with some US colleagues, security plans, a production planning session, and a misbehaving TV. My team has been at work all weekend so they have the next two days off.
0
0
1
34
I see people posting about getting more Claude or Codex accounts because they can work so fast on their product with them… and I’m sitting over here wondering who will buy their product now that everyone else is also getting more accounts to build their own versions of every good idea.
1
0
2
187
Update on the "I upgraded my X/Grok subscription and lost my Blue Checkmark debacle...". It is now day 12 since I upgraded and my profile went under review. I've messaged back and forth with support and apparently there's a bug....and they don't know how to fix it or when... yay 😂
1
0
2
137