bdan

@chetaslua I fear anthropic actually wouldn't mind dropping their subscribers too much. I fear the only thing stopping them is destroying the relatively good image they've built up against OpenAI. boy would it be a relief if they can stop supplying all those subsidized tokens...

Weird Situation for anthropic if they remove fable , opus 4.8 is so washed against SOL and super costly against grok 4.5 and muse 1.1 there will be no reason to get claude subscription anymore , only way is to launch Opus 5 asap I have tested Claude Honeycomb Anthropic research model. Early access preview. 1M context window

I like the idea, and I think the mapping between blocks and behavior as you’ve described it is already there - each sem diff links back to the exact bit of code it represents I’d be hesitant to add much beyond that. The idea behind this tool was to create the simplest, most accurate semantic representation of a code change. Having a model infer relationships outside that - like exactly which lines a test is relevant to - feels slightly out of scope and risks presenting inference as fact. Plus, the factual relationship between tests and executed lines is probably better handled by test coverage tooling

@__endgamer I feel this bottleneck too. Reviewing a generated diff gets much faster when the agent can map a block back to the decision it implements. I would want sdrev to keep links to the exact lines and tests, so “intent” stays auditable.

Converting code into intent is by far the most time and cognition demanding task I encounter now that my job primarily consists of reviewing AI generated code. I decided to get an agent to write the simplest and least annoying to use solution to this problem. A simple `sdrev "Show the changes for the last few feature X related commits"` outputs a transient browser PR-review style diff. The main difference from a typical PR is that instead of being served a code diff, each hunk is represented by a semantic explanation of what's going on. If the explanation piques your interest, you can then click it to dip into the implementation. The code translation step is largely eliminated, and you can focus on reviewing behavioral changes (i.e. what really matters). There's a 'semantic diff' json might be useful for other workflows. Uses claude/codex by default, API key support available github.com/finallyblueski…

@goodside happy to be corrected, would love to know how it actually works

@goodside im pretty sure images go through some ocr so the model will relay what that layer sees - something that it has no control over

GPT-5.6 Sol and Claude Fable 5 try to read “handwriting,” but it’s actually just meaningless scribbles I drew on my phone with my eyes closed. Sol seems to reliably hallucinate an answer. Fable admits it can’t read it, or even pushes back and says it’s not actually writing.

@ShittyTwittter @JaydenDavisNC real talk why would they expand the opus line if they have fable? opus will forever have a special place in my heart but let's not pretend it wasn't a very flawed model for coding

@JaydenDavisNC Or drop opus 5 soon, which I’m not too excited for based on sonnet 5 other they have no moat …a new grok and ChatGPT are probably coming next month. And grok is way more fun to use it’s ridiculous

Anthropic has to keep Fable 5 on the 200 plan.

@trq212 I hope you guys are ignoring the haters. Fable + your entire software suite has been bringing me so much joy as of late. Keep it going 🦀

Enjoy more Fable!

@Brobro @mralicicek @nickchapsas I have been toying with idea of buying more Max plans but I've read somewhere that Anthropic doesn't like that, and that you can get banned for it. Any truth to that?

@mralicicek @nickchapsas Used the 5 minimum accounts. Each premium which gives 6.5x usage vs 5x of a standard pro plan. I use one max 20 separately - now that I think of it using a couple 20 max plans might have overall been cheaper but I liked easily upgrading seats as needed.

I think I took this AI coding thing a bit too far this week

@wholisticguy @antirez Claude Code w/ experimental teams mode has been sufficient for my use cases

@__endgamer @antirez The tools for doing this orchestration are still pretty unimpressive though If that’s the intention, it needs to be more seemless so that fable is actually orchestrating, and minimising its own token use

There is something I don't understand about Fable: they want to go b2b with Fable available only via API at that absurd price point. I believe this is delusional (OpenAI exists) but let's buy it for a moment. But: they ask for 50$ 1M tokens with this switch-back-to-opus trigger?

@ZackKorman bro 7th of July is not that far away plz stop 😩

As long as you don't use the word cybersecurity or vulnerability, Fable will basically do whatever, including finding ways to bypass security controls. Lmfao Andy Jassy help

@fcoury @_Sagiquarius_ I think design is probably one of the harder to verify and train for things out there right now. You probably just got lucky the first time around

@_Sagiquarius_ As I said this was more of a vibe check. I gave it a UI redesign task that it would have nailed before and it did a terrible job. With that said, it was a one off test and going by vibes alone.

I played with Fable a bit when it was first launched and played a bit with it again now: am I the only one who feels like it's not even the same model?

@JerryBeller1 @Tristanrhee3 Dis u?

"Founders" = the latest word dumbed down to mean nothing. They'll be something else soon, as 99% of these new "founders" are chasing fool's gold, pretending to be experts when they do not know quality code from slop, a smoke test from a joint, or a browser test from a google search.

If everyone is a Founder, who's left to be the user?

Being on CC max plan rn be like I'm fifty percent through my weekly fable limit but we don't worry about that thank you Anthropic for letting us plebs touch the clouds

@MikeYup10 @keithedwards why don't you show us your shirtless pic bro or any newsworthy achievement for that matter

Does he think he looks muscular? Muscular dystrophic maybe. Hes anorexic. Thats not a six pack, those are his ribs, zero fat layer to protect his vital organs and ab muscle thats almost nonexistent. Hes like around 5'7 to be standing next to a female and be almost the same height. Why would he post a shirtless Pic? Lmao

This is the couple who climbed to the top of the Empire State Building

@Bartosz1944Pl @The_Banned_Vids Brother that was at the start of the war, and militia/civilians only. No Ukrainian infantryman is going to be carrying Molotovs on the frontline

well according from what i heard: ukrainian goverment start program to teach militia and soldiers to "how to make Molotov cocktail" so its posible that some soldiers are using this. even with modern RPGs, Javelins and other anti tank weapons. i checked google and they are some articles about it. so it is posible to see there a ukrainian soldier throwing molotov on tank. ofc its very rare to see this (not as much as RPG and javelin in action). but still its posible

I totally understand!

@Rainmaker1973 @grok am I being bamboozled or is this actually a fun fact

School Chalk under an electron microscope. It’s almost entirety made up of fossilized shells of ancient single-celled marine plankton called coccolithophores.

@CtrlAltDwayne Barring the fact that releasing Fable without coding abilities would be absolutely fucking stupid there are still people on my timeline and in your comments arguing that this is even remotely a possibility. x is an echo chamber of stupidity x.com/trq212/status/…

Seen a lot of people saying that Fable 5 can't be used for coding, which is not what Anthropic said at all. Some of you have skill issues around reading comprehension. It says "some" not all "some" meaning the classifier is probably sensitive depending on the task.