Christopher Little-Savage

268 posts

Christopher Little-Savage

Christopher Little-Savage

@cjols_

agent of progress

London, England Katılım Nisan 2009
144 Takip Edilen73 Takipçiler
Matt Pocock
Matt Pocock@mattpocockuk·
@dexhorthy I actually think 'Automated Check' makes most sense, since it contrasts with 'Automated Review' (for using agents as code reviewers)
English
6
0
18
6.5K
Matt Pocock
Matt Pocock@mattpocockuk·
Which term immediately makes more sense to you? (definition in the post below)
English
28
1
52
31.3K
Oluwafemi R. S
Oluwafemi R. S@itz_feminisce·
@Cloudflare But humans have to create a cloudflare account first yeah??? I get the point 😃
English
2
0
1
4.4K
Cloudflare
Cloudflare@Cloudflare·
Starting today, agents can now be Cloudflare customers. They can create a Cloudflare account, start a paid subscription, register a domain, and get back an API token to deploy code right away. cfl.re/4sY0Uxn
English
161
822
5.3K
1.6M
Christopher Little-Savage
@threepointone I'm not sure I've ever seen code displayed satisfyingly on mobile at all, even pre-agent. The problem is worse now that all we do is basically review, but it's not new
English
0
0
1
39
sunil pai
sunil pai@threepointone·
now that the basics are falling into place, the thing I'm really struggling with is UI. every "coding agent" kinda sucks on laptop/mobile, and everyone punts on it by just hiding the details "you don't need to see the code anymore". balderdash. we need to rethink it all.
sunil pai@threepointone

working on subagents / agents-as-tools for single threads, and it's so very satisfying with Think because it's just nested chat all the way down, persistence/streaming/resumption all ootb landing this week

English
17
2
65
5.8K
Adolfo Builes
Adolfo Builes@abuiles·
@threepointone do you guys have some patterns for security binding credentials/env vars in Think extensions I want to let people set credentials/secure vars per workspace and then be able to use them extensions or facets for example agent writes facet code that shows a button “send notification”, on click, it sends a post to a Google workspace or hits any other api that requires access tokens cc @jonas curious how you solved this in your current experiment
English
3
0
0
120
Christopher Little-Savage retweetledi
rita kozlov 🐀
rita kozlov 🐀@ritakozlov·
workers ai
English
7
6
128
27.2K
Christopher Little-Savage
Why are all the voice modes so much dumber than text? Is it the low-latency required to keep the conversation seamless? If so I'd rather have the option of a stilted conversation so I can make progress with my hands free
English
0
0
0
9
Christopher Little-Savage
@threepointone On the other hand I do want my agent to be accessible on multiple channels and with knowledge shared across both. Not sure how best to model - cross-channel search, or should we prepare cross-channel context? Not to mention choosing the best channel for proactive outreach
English
0
0
0
11
Christopher Little-Savage
@threepointone Hmm I'm not sure. Massive for an admin panel view or something, but the way I email and the way I slack are different and I think I want my agent to do the same (emails in Slack are already a thing and it's bad)
English
1
0
0
65
sunil pai
sunil pai@threepointone·
quick weekend hack: what if you could sync a _single_ conversation across _multiple_ messengers (and streaming / reactions / buttons etc to Just Work ™ ) I wonder how weird I could get with this. sync the same conversation with email? a phone call? There's definitely no chance to get perfect parity, but I wonder... yay agents sdk, relatively easy to get this to work.
sunil pai tweet media
English
13
0
55
7.1K
Christopher Little-Savage
@mattpocockuk have you found Grilling quality to go down with cheaper/models? Thinking of using local models for offline brainstorming on rural walks
English
1
0
0
13
Christopher Little-Savage
@Cloudflare @threepointone I'm especially curious to try this for non-code use cases to power document version control. I want to let my agents loose on arbitrary text, and then have the ability to easily review/rollback etc each chunk of changes
English
1
0
0
537
Cloudflare
Cloudflare@Cloudflare·
We’ve just launched Artifacts: Git-compatible versioned storage built for agents. cfl.re/4cJqd1n
English
12
107
653
141.2K
Mario Zechner
Mario Zechner@badlogicgames·
seeing a pattern here. friendship with CF has ended. vercel is new best friend now.
English
12
5
156
32.3K
Christopher Little-Savage
How do you combine the user's desires for the software with the domain expertise of the original author? What does backwards compatibility look like with agents that can patch breaks? What's the right API shape for exposing style and content and behaviour overrides?
English
0
0
0
43
Christopher Little-Savage
I'm excited for personal software as a user, but I'm even more excited by it as a problem space. So many unanswered questions!
English
1
0
1
57
Christopher Little-Savage retweetledi
Cloudflare
Cloudflare@Cloudflare·
Announcing a preview of the next edition of the Agents SDK — from lightweight primitives to a batteries-included platform for AI agents that think, act, and persist: cfl.re/485sEZr
English
7
18
187
57.5K
Christopher Little-Savage
Like when I have actual need for an answer, I don't want to spend turns coaching it. And when I do, I don't trust that it's answering from it's true configuration rather than just based on the recent conversational context
English
0
0
0
9
Christopher Little-Savage
I still haven't cracked openclaw. Feels muddled to have same interface for using the agent as for defining the agent
English
1
0
0
18
Christopher Little-Savage retweetledi
claire vo 🖤
claire vo 🖤@clairevo·
Agentic parenting is so unnatural, completely out of touch, "real" moms don't need all this, they keep it simple with the basics: - mom calendar, dad calendar, school calendar x however many kids, sports calendars x however many kids - mom email, dad email, school app x however many kids, whatsapp group for each classroom and sport / activity - shopping list in note or in app or on refrigerator - meal plans in their brain or on a whiteboard - pile of mail on the countertop - group text with mom friends of various sorts setting up playdates and coordinating carpools - amazon wishlists for classrooms + school supplies - google spreadsheets for snack signup for class parties and game days - cobbled together camp spreadsheets + hours dedicated to signing up on various activity platforms - homework nights bouncing between kiddos trying to remember your 8th grade geometry - crumpled growth charts at the bottom of your diaper bag from the pediatrician visits - waiting on hold with the pediatrician to find out if they can email --no, not fax-- vax records to the new school - looking at your kids pants going wait did you grow out of your clothes again, putting said pants in that "I swear i'll get a donation run in" bag that's been sitting in the garage, forgetting the bag, forgetting to order new pants - starting your kid on a delightful educational yt channel, looking away for 7 minutes, and accidentally exposing them to bizarro AI slop - googling "what fruit is my baby" when pregnant - carrying around a giant binder of printouts appointment to appointment if (god forbid) your child has special needs or a complex health issue - taking pictures on your phone you never have time to print - printing "kindergarden handwriting printables" off etsy or pinterest when you notice wobbly writing - scrolling doctor becky videos on insta when you notice your tween is sassy "Just be a mom." FFS
a16z@a16z

Jesse Genet on Agentic Parenting Jesse Genet joins a16z's Sarah Wang and Katherine Boyle to discuss her journey from founder to parent, how she's using agents in her household, and how AI could transform parenting for the better. 00:00 YC founder turned homeschool mom 03:00 Discovering Claude Code and agentic building 06:00 Building while homeschooling 4 kids under 5 11:00 How AI generates personalized lesson plans and logs progress 18:00 Jesse's 11-agents 27:05 Agent tech stack deep dive 33:56 How agents improve daily life 40:04 Letting kids interact with AI: values, risks, and the future of parenting @jessegenet @KTmBoyle @sarahdingwang

English
45
19
466
75.2K
Christopher Little-Savage retweetledi
Ryan Florence
Ryan Florence@ryanflorence·
I'm not sure anybody has AI coding figured out, but some sure think they do I've experienced the entire spectrum from "my job is cooked" to "agents are chopped, doing this myself" I see their tweets and think "ah yes, I remember being on that point in the cycle last month"
English
31
20
389
17K