rsanheim
Okay, checked, it was March of this year: https://futuresearch.ai/blog/litellm-pypi-supply-chain-attac.... What even is time anymore in the llm age?
However, our world is not that world. Your agents may be happy in their tiny walled kingdom of toil and ineffectiveness, but you and your users will not. Poorly drawn service boundaries will drag you down more than almost any other architectural mistake.
If there is one universal amongst organizations, its that they love walls and silos. Be wary of putting them up ahead of time, cuz tearing them down once established is nearly impossible.
edit: replaced false prophet site with wiki link for the original one and true timecube.com...lost to the sands of ...time.
"load-bearing" I have never heard used for programming before Opus, and its incredibly annoying and over-used.
the closest thing i've found are tooks like https://github.com/raylabshq/gitea-mirror
No.
I understand how we got here, where many experienced programmers, managers, and bloggers only know capital-A Agile as the watered down version sold via certifications, crummy medium posts, and atlassian flavored kanban boards. But that isn't agile.
I can’t even with the pitch into spec driven development as some sort of high watermark of software methodology.
If this isn’t something to at least root for, in the sense of a small team, novel product, serving a real need, then I dunno what is. You can use jj or tangled and still appreciate improvements to git and vcs on the web in general. Competition amongst many players is a good thing, even if you don’t believe in this one particular vision.
Heaven forbid it isn’t 100M going to a YC alum for yet another AI funding raise.
The continual recall/ai push from Microsoft has not helped at all and is pretty gross. There is a way to do a “recall” style thing that some folks will really want if they can trust it. The msft approach has been the opposite of that.
And it really comes down to $MSFT. If the stock keeps dropping, how long do you think any real commitment to “quality” for a boring, low(no?) revenue product will last? Very little when the ad/partner revenue really starts flowing for “ai focused metrics” that can directly tie to windows surveillance (ie recall).
My limits are now many of the same things that are have always been core to software dev, but are now even more obvious:
- what is the thing we are building? What is the core product or bug fix or feature?
- what are we _not_ building? What do we not care about?
- do I understand the code enough to guide design and architecture?
- can I guide dev and make good choices when it’s far outside my expertise but I know enough to “smell” when things are going off the rails
It’s a weird time
The CLIs I’ve seen agents struggle with are those that wrap an enormous, unwieldy, poorly designed API under one namespace. All of Google Workspace apis, for example.
The original name of his ai assistant tool was 'clawdbot' until Anthropic C&D'ed him. All the examples and blog posts walking thru new user setup on a mac mini or VPS were assuming a claude code max account.
I know he uses many llms for his actual software dev.. - right tool for the job. But the origins of openclaw seem to me more rooted in claude code than codex.
Which does give the whole story an interesting angle when you consider the safety/alignment angle that Anthropic pledges to (publicly) and OpenAI pretty much ignores (publicly). Which is ironic, as configuring codex cli to 'full yolo mode' feels more burdensome and scary than in Claude Code. But I'm pretty sure that speaks more to eng/product decisions, and not CEO & biz strategy choices.
This is not new. Anthropic has raised these concerns in their system cards for previous versions of Opus/Sonnet. Maybe in slightly more dryer terms, and buried in a 100+ page PDF, but they have raised the risk of either
a) a small group of bad actors w/ access to frontier models, technical know-how (both 'llm/ai how to bypass restrictions' and making and sourcing weapons) to turn that into dirty bombs / small nuclear devices and where to deploy them. b) the bigger, more scifi threat, of a fleet of agents going rogue, maybe on orders of a nation state, to do the same
I think option a is much more frightening and likely. option b makes for better scifi thrillers, and still could happen in 5-30ish(??) years.
I recognize 100% that a tool to manage ai agents with long term context tracking is going to be a big thing. Many folks have written versions of this already. But mashing together the complexity of k8s with a hodge podge of lotr and mad max references is not it.
Its like the complexity of J2EE combined with AI-fueled solipsim and a microdosing mushroom regime gone off the rails. What even are all the layers of abstractions here? and to build what? What actual apps or systems has this thing built? AFAICT it has built gas town, and nothing else. Not surprising that it has eaten its own tail.
The amount of jargon, ai art, pop culture references, and excessive complexity going on here is truly amazing, and I would assume its satire if I didn't know Yegge's style and previous writings. Its like someone looked at the amount of overlapping and confusing tools Anthropic has released around Claude Code, and said "hold my beer, hand me 3 red bulls and a shot of espresso, I can top that!".
I do think a friend of mine nailed it though with this quote: "This whole "I'm using agents to write so much software" building-in-public trend, but without actually showing what they built, reminds me of the people selling courses on stock trading or drop shipping."
The amount of get-rich quick schemes around any new tech are boundless. As yegge himself points out in the post towards the end, you'd be surprised what you can pull off with a ridiculous blog post, big-tech reputation, and excessive LOC dev-tools in a hype-driven market. How could it be wrong if it aligns so closely with so many CEOs dreams?
even more so when considered in the context of dev 'remote workstations'. I benchmarked perf on AWS instances that was at least 5x slower than an average m1 macbook, and cost hundreds of dollars a dev per month (easily), and the macbook was a sunk cost!
Sony specifically targets the reference with their final calibration on their top TVs, assuming you are in Cinema or Dolby Vision mode, or whatever they call it this year.
Sometimes, you might really know better, and it doesn't matter. You build the thing with the wrong tools, with a crummy framework, with a product at the end that will probably not succeed. But that is okay, hopefully you learn something and your team and your org learn something.
And if not, that is okay, sometimes its just a job and you need a paycheck and a place to be from 9 to 5.
In any case, how is any of this better than a CLI? CLIs have the same access models and tradeoffs, and a persistent agent will plumb the depths of your file system and environment to find a token to do a thing if your prompt was “do a thing, use tool/mcp/cli”.
So where is this encapsulation benefit?
And we don't really have a simpler alternative in sight...at least in the java days there was the disgust and reaction via struts, spring, EJB3+, and of course other languages and communities.
Not sure how we exactly we got into such an over-engineered mono-culture in terms of operations and monitoring and deployment for 80%+ of the industry (k8s + graf/loki/tempo + endless supporting tools or flavors), but it is really a sad state.
Then you have endless implementations handling bits and pieces of various parts of the spec, and of course you have the tools to actually ingest and analyze and report on them.
It was a fun little post that felt accurate (ie confirmed my own biases ;)) about the current state of LLM models in a silly, but real, use-case.
The continual drive to out "llm written" articles feels a bit silly to me at this point. They are now part of the tools and tech we use, for better or worse. And to be clear, I think in a lot of cases it leans towards 'worse'.
But do you question if a video or photo was made with digital editing or filters or 'ai' tools (many of which we've had for years, just under different names) ? Do you worry about what tech was used in making your favorite album or song?
I get it, LLMs make it easy to produce trash content, but this is not a new problem. If you see trash, call it out as trash on its flaws, not on a presumption of how it was made.