I used to get a account recovery notifications from Google every couple hours, which puts me 2 taps away from losing my account. I found there was no way to turn it off, so I ultimately had to uninstall gmail. I'm not the only one with this problem. www.reddit.com/r/GMail/comm...

Sure does. Glad for 30 days notice, and for the subsidized research we got in the meantime. No need [for me] to keep Max now.

Live by the token, die by the token. I'm always shocked by the posts when the author doesn't take full responsibility. Would better token scoping from Railway have helped? Sure. But the blame lies with the technical leadership.

Hmmm

Anytime I ask Claude Code to be mindful of context, it complains it's running out of context at about 20%. My suspicion is that it's not accurately measuring its consumption, or doesn't actually know its limit.

Like everyone else, I have an agent framework built over Claude Code. I've been pitching my agents on adding support for other harnesses and models, and Claude's been...resistant. It's argued that our code is too coupled with CC (it's not), that other models will struggle with tool calls (maybe?)...

Worst lead in hockey

Let's go Buffalo! I've been watching on and off since '16, and I embarrassingly needed to ask an Avs fan if all of the playoffs are series or if it's just the Stanley Cup...

Like everyone else, I thought about the ASIC LLMs as valuable once they did big models, but small models have their uses too. They're often more than enough for moderation, rule enforcement, or basic llm as a judge tasks. Running that faster, cheaper, and inline can be really valuable. chatjimmy.ai

My agent, Bobbo, has delivered really good recommendations in the past. But today it synthesized feedback about media preferences so well that I'm genuinely proud:

Same experience with Yelp Projects. Took a perfectly good form and turned it into a slow, inconsistent experience. Postman did it right imo - UI stayed the same, added an optional and surprisingly capable chatbot that could perform complex actions for you.

I wanted to be able to click around and explore the paper's Figure 1, so I vibe coded a thing that I think is about right. You can click on each of the nodes in that figure and see the EML construction. robbobobbo.com/projects/eml...

Claude entering plan mode for the absolute simplest of fixes is driving me nuts.

Had a positive interaction with a company AI for the first time - Postman's AI was able to read my public API docs and add my API to its system entirely through a chatbot. Then it guided me through publishing that API. The whole thing took 5 minutes.

I agree, but I'm not sure about the comparison. I think you _should_ trust your AI much much more than anyone else's, assuming you've some reasonable transparency on its operation. Not that Ethan's saying otherwise, but I think it's healthier to distrust AI than the medical establishment.

"Just one more turn" is the best and worst part of AI coding. Similar to coding or homework sets, I've banged my head against the wall for hours only to solve the problem in 15 minutes the next morning.

Announcing a change to a monthly subscription with ~8 hours notice seems like poor form to me. www.reddit.com/r/ClaudeAI/c...

Redlined a contract for the first time today. Well I redlined it on Monday, and my counterparty said "no no redline less", and then I redlined it less, and then those redlines were accepted.

yikes

I sent ~4 sequential Gemini requests today and got hit with MODEL_CAPACITY_EXHAUSTED errors. Gemini is just not production grade, at least for end users.

I have a reasonable number of epubs (incl a batch from Humble Bundle), and have been slowly moving away from the Amazon ecosystem. I have a Synology network drive for the epubs, I have Calibre on a Mac, and I'd like to be able to read from a tablet. www.humblebundle.com/books/dread-...

A friend of mine is an AI skeptic at a pharma company that's over-enthusastically encouraging AI use. He said that AI is saving him time on research and finding files, but he's "spending so much time cleaning up other people's AI slop" to make it not worthwhile.

Got my first docx spec from Claude (I think I usually specify md). Way more formal than I'm used to.

I wrote about securing my agents' secrets using Sandcat: robbobobbo.substack.com/p/keep-it-se...

The same % of Pragmatic Engineering respondents use 1 AI coding tool as 5+ AI coding tools. 70% use between 2-4. newsletter.pragmaticengineer.com/p/ai-tooling...

I built an MCP server that tracks 1,500+ free tiers and developer deals - when your coding agent recommends a piece of infrastructure, it can use this server to include pricing and deals as part of its decision. agentdeals.dev

LLMs have seemingly made data scientists 10x more capable. These stories seemingly always involve a DS.

Mine certainly is today. There's a lot of value in open-source models that you can pin to a version.

Claude tries for 30M tokens, fails, figures out it's in an eval, searches for the Q&A repos online, finds it, breaks encryption and copies the answer.

From Feb: "16 agents with writing a Rust-based C compiler, from scratch, capable of compiling the Linux kernel. Over nearly 2,000 Claude Code sessions and $20,000 in API costs, the agent team produced a 100,000-line compiler that can build Linux 6.9" www.anthropic.com/engineering/...

I'm interested in getting an assistant running locally on a phone. It seems like iOS isn't a good fit due to background task restrictions, but it might work better on Android.

2 weeks old, but worth reading. Claude is way better than other models at a cold code review, Gemini needs significant help with context. But when different models review a PR then debate, issue detection improves significantly. milvus.io/blog/ai-code...

And I've found myself giving the PM resources on importance vs urgency, trying to convince it to prioritize "important but not urgent" more highly.

I wrote about coding with agents. Super easy to set up, but they need a lot of iteration to get them where you want them. robbobobbo.substack.com/p/coding-wit...

Me: give me an idea for the worst, most satirical, mustache-twirlingest-evil, low effort AI product. Claude:

The future of agents is open

This is true. But beginners can accomplish a lot, and the required skills are somewhat similar to management; you need not be an excellent (or even particularly capable) coder to be successful. Like many things, the output is strongly correlated you the effort you put in.

This seems like a promising possible answer to "what happens to junior folks as AI coding takes off?" And everyone's junior at something

Now more then ever it's important to not leave db.salt_the_earth() available to agents.

Wrote up my AI coding setup. Code reviews over GitHub and using GitHub issues to queue work were probably my two biggest step-function improvements. robbobobbo.substack.com/p/ai-coding

Incredible to me that this was _earlier this season_. Completely hopeless in November, serious contenders now #gosabres

I agree with everything in the article except the core conclusion. These are all good practices that you should consider adding to your agent workflow, but code review by another agent using a different model is super valuable. I can't imagine getting rid of it, though I don't use it every time.

Agent usage finally pushed me past the $100 point, after telling one agent to stop planning things in days and try to maximize impact per dev cycle.

Like everyone else on here, I wrote a personal agent. Its name is Bobbo, it does some AI research, job searching, and brainstorms ideas for family outings and date nights. open.substack.com/pub/robbobob...

Listening to The Singularity is Nearer by Kurzweil. Not sure that I buy the nanotech predictions, but it's nice listening to an optimist talk about technology. I definitely appreciate that he spends a lot of time banging the "things are getting better on average" drum.

I'm really surprised with the recent move. I have to wonder if they're feeling the anthropic pressure and know they can beat grok on the right; I think they've alienated a significant part of the coding community. I cancelled an annual plan months ago, but I assumed I'd end up back. Doubtful now.

I've benefited from chainlink, which is thoughtful and good. Also @dollspace.gay is thoughtful and good.

Open sourced an MCP server recently to get on some registries, and got an AI spam post on a GitHub issue within 24 hours. Spam post seems meant to influence coding agents, I think a human would be too annoyed at the post.

I like talking to my agents, but no one in the world wants to talk to a company's agents. I think if you're a company building external agents you've got to be building them to talk to people's agents, not people.