viccis

I'm, by contrast, an able bodied person with functioning eyes and had no issues.

https://news.ycombinator.com/item?id=48766580

Here's an example of one for a major result earlier this year: https://cdn.openai.com/pdf/1625eff6-5ac1-40d8-b1db-5d5cf925d...

Who knows what "rewritten" means, but you can sort of see how it progresses.

If the human solved a significant open problem without relying on AI, don't you think they'd have claimed the credit for themselves?

You'd likely make a lot more money if Anthropic paid you a couple million to release it as a Fable discovery. For that matter, if "I solved it" bragging rights and CV candy is that important, then why not just publish the counterexample without mentioning Fable?

I was initially a bit skeptical, as it seems very convenient that this discovery came from an employee of the company that was able to take credit for it. When I saw no real prompt was published, that was even more suspicious. Some recent mathematical discoveries included the prompt and session [1]. Why not include the prompt? That and the thought the model used could be incredibly valuable information for the field of mathematics.

I'm not even saying Fable didn't do it, and my attitude might be different with other companies, but Anthropic has so much obvious nonsense marketing going on (their AI welfare experts for example), that I think the burden of proof should be expected to be on them here.

1: https://chatgpt.com/share/6a55aa50-b484-83ea-85c0-c7e7b4bda4...

Mostly because the person I was replying to has commented about using it to write code.

If you're using it for other purposes, then I give you permission to ignore my comment; there's no reason to descend into name calling.

So the price is rising and you have no choice but to keep paying more and more.

You can also just write code like you did a year or two ago.

"Supporting audiobooks" involves more than just playing audio files, unless you want to ignore the basic set of features associated with audiobooks as their own medium.

You're inventing a person to get mad at. I don't see any evidence that this founder holds those kind of views.

Easily the biggest delta between a good one and a bad one I've ever encountered. It's so easy to test too. Just ask a product manager if they can give a talk on <some product they own> in 2 hours. If they respond by slacking all of the engineers that work under them to stop what they're doing and give them the info to give that talk, fire them. If they are ready to talk right now then give them a raise.

China's doing it

From what I can tell, I have a similar workflow. I get Fable 5 / Sol on High to talk to me about requirements until it's ready to design. I then have it break the design into tickets. I use kata, an agent-oriented issue tracker. If it ever starts to get bloated, I'll just make my own as it doesn't need too many features. Anyway, I tell it to include sufficient context in each such ticket to be picked up by a new implementation agent. Explicit user stories, Cucumber style acceptance criteria, etc.

When it's done, I switch to a smaller model and tell it to start a /goal of calling `kata ready` to get tickets ready to be worked. Work one ticket on each goal iteration, committing changes when done. Stop when all the tickets are either closed out or are blocked on actions from me.

It works fantastically. I can get entire (relatively straightforward) iOS apps done in under my $20/month five hour session window.

I get even better results if I do the QA session with the frontier model with two output artifacts: an implementation spec and a design prompt for Claude Design. I push the design prompt through Claude Design, tweak the results, and get a design spec. When I have the frontier model do the planning, I have it read the implementation spec from before, along with the design spec's overview file.

When I do it this way, I've done A/B tests between Fable 5 and Sol, and the apps they wrote were basically identical. It's so much cheaper than the "let loose the subagent fleet!" form of context management.

I haven't even added any kind of frontier model validation cycle into this loop yet. Everyone keeps talking about a subagent flow in which the big boy reviews the work of the drones, but I've found that if it encodes its acceptance criteria well enough, they do a satisfactory job of it themselves.

I don't see how that particular political expression is a risk to their VPN's stance. This is a lot of nothing; would have been better to ignore the screeching from Bluesky.

Not really, this looks more like pre-1990s leftism that you see from people like Caesar Chavez. Broadly redistributive policies towards the working class and a strict curbing of immigration as a source of exploited labor underclass for corporations to fatten their margins. There's a reason Bernie called open borders a "Koch brothers proposal".

The bottleneck has always been on the product side for 99% of cases. For the 1% where it's engineering, LLMs don't help as much.

We're rediscovering the "architect" job role a decade or so after switching to agile processes and staff engineers largely replaced it. God help us when they rediscover UML for agents.

I started pre-1.8, in the C-key era. Haven't used it in many years so looking at it every now and then is genuinely shocking to me how much it's progressed. I remember a family member who worked at a triple A game dev studio giving me a tour and when I talked to the 3D modeler guys there and told them I used Blender they stifled a chuckle. To be fair, it was a very rudimentary tool back with UI flows invented by and for aliens. And now it's used all over the place in both amateur and professional settings.

As one of them, I'm just gonna buy another $20 subscription to OpenAI and use Sol while that's available. Why on earth would I do anything else? Fable is not magical enough to pay for credits for it with the current competition.

At some point, narcissistic injuries in response to "rtfm" or "stfw" got so prominent that just telling people asking others to spend their time explaining things that were 5 second google searches away to google it became a faux pas

Makes me nostalgic for the old days (2000s and early 2010s) when everything was WordWord. Facebook, Instagram, Myspace, Dropbox, etc. Now it's just normal words without vowels. The Grindrification of naming.

I love this author's article about saving half a million dollars with a click from a while back. Nikhil, if you're reading this, I have a decent war story about a similar situation (I was lucky enough that there were two such things, so I saved a full million a year and was still denied a $15k raise) and so I was really entertained at your post from a few years back. There are actually a lot of lessons to be learned about corporate politics there, about how you can save someone a million a year in perpetuity and promise to do that again next year (which I could have!) and see them still refuse to pay you a single extra cent.

Looks like no Fable for Pro. Will likely be cancelling my Pro account.

When you've earned your opinions about architecture and code quality the hard way, they feel less like textbook rules and more like scar tissue.

I don't think it's common for any compsci programs to (competently at least) teach architecture and code quality.

The honest truth is that in the last few months, there have been days when I have spent close to two full days writing a plan for an LLM to execute: obsessively clarifying, specifying, re-specifying, only to have it still do something inexplicably stupid.

It's because LLMs are actually taking us back in time to the pre-agile days where there was a career path (architect) that involved almost nothing but painstaking spec authoring and endless meetings to review and course correct the work of the engineers whose job was to implement what you designed as closely as possible. I have to emphasize that this was a different career path than what we think of as a senior engineer today. Not everyone likes this.

Yeah, pseudpocalypse would be Substack going out of business or something similar.

That's the saving grace of it for me tbh

AI art shows the philistines for who they are. It's a giant decoy to lure them away from spaces where people are making art of more substance.

I feel the same way about things like the absolutely miserable quality of Marvel, Star Wars, etc. stuff leading to films being increasingly polarized between 60 IQ blockbusters and more interesting small projects. It's what kicked off the 70s arthouse cinema in the US, and I'm hoping we get a similar renaissance soon with the success of all of the small and difference projects lately.

doing concerts where the attendees are just other musicians

If this is how jazz musicians actually made money, they'd have all died of starvation at this point. Every jazz musician that's not a top 20 or so act makes most of their money playing whatever kind of music people pay them to make, whether it's church music, session musician gigs, etc. The entire reason they started doing (and still do!) straightahead jams is that they were tired of playing schlock all day and wanted to spice it up with and for other people tired of playing schlock.

Andy Warhol

Given that the man made his living doing stunts on top of the corpse of modernity, I'm not surprised he would love anti-human stuff like AI music. He famously said he wished he could be a machine. I don't think appealing to his nihilistic embrace of the spectacle is persuasive.

We should also be mindful of how much these tools break down the "be careful and thoughtful" barriers in favor of more and more convenience.

Yeah I had it remake my favorite TI-89 graphing calculator game in Python and it one shotted it perfectly.

If only a cynical Frenchman had written a book critiquing peoples' tendency towards simulating things that don't exist.

Neocities is cool, but the medium is the message and we've generally moved on from this (treasured!) past. Any attempt to replicate it tends to wind up hyperreal and forced.

Same for me. I never use them. I use Fable on highest effort to plan things and then record the plan in tickets. I use Kata, which is CLI and agent oriented, but I suppose Jira or other systems would work too. I tell it to put enough context in each ticket to on-board a fresh coding agent to implement it. Then I just do /goal, telling to to run `kata ready` to get new tickets to work and continue until they're all closed according to acceptance criteria or until they're blocked on actions from me. I need to play around with getting it to switch to smaller models (or spawning 1 subagent) to do ticket implementation and then auto compact after each. Either way, it results in really easy workflows and uses very few tokens compared to the built in subagent flows that doing this completely avoids.

No, in fact, most of us could not have just shown up and bid on it with any expectation of a meaningful outcome.