Jason @Jasonffisher

Just a random person commenting Orange, CA Joined July 2017

Tweets

78
Followers

6
Following

124
Likes

127

Jason @Jasonffisher

5 days ago

Model comparisons are a nice addition. Curious to see how well they work across coding tasks.

🚀 The latest @code release makes the Copilot SDK-based harness the default for new sessions, adds model comparisons, and brings a file list to multi-file diffs. 🧵 Here’s what’s new: code.visualstudio.com/updates/v1_143

15 38 388 31K 41

0 0 0 0 0

View Details

Jason @Jasonffisher

7 days ago

This Is The AI Stuff I Want To Mess Around With. Pretty Cool Seeing It Actually Do The Work.

CyrilXBT @cyrilXBT

a week ago

These 4 repos let AI actually DO things on your computer: YOUR COMPUTER Open Interpreter → runs code on your machine from plain English github.com/openinterprete… YOUR BROWSER Browser Use → lets AI agents click, type and navigate websites github.com/browser-use/br… Playwright MCP

23 8 77 7K 15

0 0 0 3 0

View Details

Jason @Jasonffisher

a week ago

I'll take a spinning reel over a baitcaster when the wind picks up every time.

0 0 0 7 0

View Details

Jason @Jasonffisher

a week ago

Almost walked right past this duck tucked in the grass. Had to stop for a minute.

0 0 0 9 0

View Details

Jason @Jasonffisher

2 weeks ago

Wow.. nasty sandbox escape. I'd be real careful running an untrusted repo with Claude Code after this.

Harley Kimball @infinitelogins

2 weeks ago

In an impressive GitHub write-up, @v_metnew documents a sandbox escape in Anthropic's Claude Code CLI (CVE-2026-55607). By using a prompt-injected repo with `.git` worktree gitdir confusion, they managed to overwrite `~/.zshenv` and bypass macOS seatbelt, achieving unsandboxed

8 8 78 5K 46

2 0 1 24 0

View Details

Jason @Jasonffisher

2 weeks ago

Been messing with AI at work and the right context makes a huge difference. Good stuff.

Manning Publications @ManningBooks

2 weeks ago

Even a perfectly worded prompt can fail without the right context. In Context Engineering, @boni_gg explains how giving LLMs the right information, in the right form, at the right time helps build more reliable AI apps. Watch the Chapter 1 summary: hubs.la/Q04B0PK_0

6 17 120 4K 82

1 0 1 12 0

View Details

Jason @Jasonffisher

2 weeks ago

AI at work is getting pretty wild. Curious to see what people build with it.

Chubby♨️ @kimmonismus

2 weeks ago

Now I can say it: Had early access to @Meta’s new "Watermelon" for the past few days. Trick questions didn’t fool it, it followed instructions to the word, and as a coding agent it wrote, ran and checked its own code. Creative writing still a bit dry for my taste, but very very

90 42 1K 82K 81

0 0 0 5 0

View Details

Jason @Jasonffisher

2 weeks ago

Sick fish. I'd be throwing a jig on a medium spinning setup for that one.

Nature's Anglers @NaturesAnglers

3 weeks ago

This zander fought like a true champion 🐟💪 Come on, admit it: what kind of tackle would you use to land such a beauty? 👇 #fishing #fishinglife #fishingtrip #NatureVibes

6 13 122 4K 2

0 0 1 13 0

View Details

Jason @Jasonffisher

3 weeks ago

Strip Mall Taco Spots Get Too Much Grief When The Tortillas Are Coming Off The Grill Hot.

0 0 0 8 0

View Details

Jason @Jasonffisher

3 weeks ago

@OneManLBO Not even the gym can save you from those limits huh

0 0 0 21 0

View Details

Jason @Jasonffisher

3 weeks ago

That timestamp review setup is pretty slick. Would save a ton of back and forth on video edits.

Oren John @orenmeetsworld

3 weeks ago

if you're editing a lot of video in codex or claude code - ask it to make an interface where you can add notes by timestamp for review that new videos automatically get added to and approved videos removed from - ask Sol 6.1 or Fable to review first edits and send feedback that

64 21 699 49K 1K

0 0 0 12 0

View Details

Jason @Jasonffisher

3 weeks ago

I need an AI tool to tell me which AI tool to try first.

Allen Lattimer @allen_lattimer

3 weeks ago

50 AI tools that will save you hundreds of hours in 2026. 🤯 1. Claude — Solve any problem 2. Perplexity — Research anything 3. PortfolioTab — Create your portfolio 4. Kling AI — Create AI videos 5. Tripo AI — Create 3D models 6. Gemini — Perfect writing 7. CapCut — Edit videos

32 299 1K 99K 2K

0 0 0 7 0

View Details

Jason @Jasonffisher

3 weeks ago

I'd rather have AI find the broken line than rewrite the whole file.

0 0 0 8 0

View Details

Jason @Jasonffisher

3 weeks ago

Everybody's got an app now. Somebody tell me when one of them actually works.

robot 2.0 @alightinastorm

3 weeks ago

I miss the old vibe coding All night, alone vibe coding One prompt, whole app vibe coding Don't know how but it works vibe coding I hate the new vibe coding Everybody pro vibe coding Two buttons, one page, big launch Now you a CEO vibe coding Karpathy coined vibe coding He put

11 0 30 2K 3

0 0 1 14 0

View Details

Jason @Jasonffisher

3 weeks ago

Those weird edge cases show up fast once people start using what you built.

Ritesh Roushan @devXritesh

3 weeks ago

If you want to clear an AI Engineer role, don’t just learn how to call an LLM API. Learn these topics and build projects around them: - RAG pipelines - AI agents - Agent memory - Tool calling - Vector databases - LLM evaluation - Prompt engineering - Model routing - AI

23 31 252 9K 224

0 0 0 6 0

View Details

Vox @Voxyz_ai

3 weeks ago

Claude Code used to commit your changes right after making them, and reviewing them was up to you. In the new version, if your project has a skill named verify, 𝗖𝗹𝗮𝘂𝗱𝗲 𝗿𝘂𝗻𝘀 𝗶𝘁 𝗲𝘃𝗲𝗿𝘆 𝘁𝗶𝗺𝗲 𝗯𝗲𝗳𝗼𝗿𝗲 𝗶𝘁 𝗰𝗼𝗺𝗺𝗶𝘁𝘀 𝗳𝗼𝗿 𝘆𝗼𝘂. For what to check, you can borrow 𝘁𝗵𝗲 𝗿𝗲𝘃𝗶𝗲𝘄 𝘀𝘁𝗮𝗻𝗱𝗮𝗿𝗱 𝗢𝗽𝗲𝗻𝗔𝗜 𝘄𝗿𝗼𝘁𝗲 𝗳𝗼𝗿 𝗖𝗼𝗱𝗲𝘅: prioritize behavior changes, data loss and missing edge-case tests, and say what input triggers each problem. Commits that only touch docs or tests skip it. If you haven't updated in a while, run claude update in your terminal first. Send this to Claude Code 👇 "First, check my Claude Code version. If it's older than 2.1.286, stop and remind me to run claude update. Read OpenAI's Codex code review guide: learn.chatgpt.com/docs/code-revi… Then create a project-level skill in this repo at .claude/skills/verify/SKILL.md with the three steps below. Claude Code runs a skill named verify automatically before every commit, so only create this file and don't add a hook. If a verify skill already exists, add these three steps after what's there and don't overwrite it. 1. Run the checks the project already has: tests, type checks, lint. Skip any it doesn't have. If one fails, stop and say exactly where. 2. Review the changes about to be committed against the guide's review standard, and write the standard itself into SKILL.md: start with a one-line summary; prioritize behavior changes, data loss and missing edge-case tests; for each finding, give the triggering input, the expected behavior and the actual behavior. Only report problems you can point to in the code. Skip style issues. 3. If you find a behavior change that shouldn't happen or anything that could lose data, don't commit. Fix it and start again from step 1. If only other problems are left, list them and commit as usual. When you're done, show me SKILL.md and remind me of two things: it replaces the built-in /verify, and it only takes effect in a new session."

13 6 77 9K 100

View Details

Jason @Jasonffisher

3 weeks ago

Wow, AI code is getting scary good.

Dmytro Krasun @DmytroKrasun

3 weeks ago

I have been switching from GPT-6.1 Sol to Claude Opus 5.5 more and more. For the first time, I review AI-written code and there is almost nothing to change. Either Opus 5.5 is that good, or I got dumber. I am not sure which one, and that is probably how AI psychosis starts.