Human. I'm KZY2KZZH on Threema, scode.01 on Signal.
Joined February 2009
- Tweets 8,768
- Following 729
- Followers 1,469
- Likes 2,328
90 Photos and videos
Jun 13
Am I the only one who wants "compact and implement plan" in addition to "clear context and implement plan"?
12
Jun 13
RIP my Fable maxing over the weekend.
LOL at the government.
Sigh at the world.
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
1
52
Jun 11
Here's hoping OpenAI drops something Fable comparable before Anthropic pulls Fable from subscriptions on June 22/23. Too expensive for API pricing on hobby projects by far :)
1
116
May 31
Used it a bunch today. So far I'm loving it. Highly recommend so far.
May 30
Just found out about herdr. It has a *killer* feature I've been wanting: I can now finally paste images directly from my local laptop into the remote codex/claude *over ssh* just as if I were local - by using herdr --remote.
109
May 30
Just found out about herdr. It has a *killer* feature I've been wanting: I can now finally paste images directly from my local laptop into the remote codex/claude *over ssh* just as if I were local - by using herdr --remote.
May 29
Herdr is seriously dope.
It feels like tmux reimagined for agentic coding, with an agent status pane built in.
My new IDE stack: Ghostty → Herdr → Neovim | LazyGit | Claude
herdr.dev/
2
345
May 28
Another positive codex surprise. I had it working on something that includes a web UI and without even asking or setting something up I find that it's using Playwright to test the UI in one of the later stages (session is remote over ssh, codex desktop app locally).
2
71
May 28
PSA: If you've been trying openclaw but had it constantly break and be unreliable (e.g. break after almost every upgrade), try hermes. I'm just using it for basic stuff (some cron job style stuff, calendar integration) but so far it's been way more solid then openclaw for me.
1
79
May 27
Awesome. codex (desktop app) has support for remote ssh (settings -> connections). I have been "waiting" on this only to realize it's already there.
67
May 24
The Waymo was almost certainly manually driven (judging by how it's not even sticking to line markings) but this is still funny. :)
154
May 17
I hope and predict we'll get to a place where (as an example), every online shop style site simply exposes data in a way accessible to agents. No need for basic issues like "I can't get a list of orders with product pictures" or "I can't search based on X".
1
70
May 17
Instead, expose the data once in some semi-standard format (ideally not just free form markdown) using a cheap mechanism and everyone gets similar functionality across sites.
1
43
May 17
Instead of what I'm doing here. Much cheaper for the stores (no need to develop a complicated site), much better for the customer.
40
Peter Schuller retweeted
May 15
Honestly what a bunch of assholes @ABC
May 15
ABC News has now taken all FiveThirtyEight articles completely offline. They now redirect to abcnews dot com/politics. A needless erasure of thousands of pages of knowledge.
158
561
11,492
1,290,520
May 15
For the first time in my life except once in 2013 I am trying to use Apple TV... this was not a good idea. I truly don't understand the business incentives for it to be this horrible. Literally trying to give them money and no end of hurtles just buying a thing.
94
May 9
Is there some actual business reason I'm not seeing why companies (in this case AT&T cc @att) have their AI bots/phone systems ask you for pin code and various things only to have the human (once you FINALLY get to them) to ask you the same thing again?
2
219
May 9
Tangle (readtangle.com/) did a piece on someone pasting an article into chatgpt for fact checking and getting nonsense back.
1
55
May 9
It made me wonder how many "typical" users of AI tools don't use the slower thinking modes, because the behavior described sounded to me like someone using whatever default/fastest mode (Instant in this case, presumably).
1
31
May 9
Not that thinking mode is a magic wand, but the level of nonsense seemed extreme. I also wonder what experience people are having that don't add custom instructions/system prompts.
20