// Lofty

Heddle Founder

[// Lofty @loftyPuma]

Exactly! With Heddle one developer can still own the result while being able to attribute code to other entities. That can nest too so agents can blame subagents for work. But at the end of the day, the human still has accountability for it.
The anti-drift stuff came about as a side effect. I needed to attack more info than I wanted to bury in comments and a maze of MD files was just drifting every 3 days. Now my context lives in the code and whenever that code is read or updated the agent automatically gets the context and Heddle raises that as a possibly stale annotation. Now you know exactly when your knowledge might be drifting by default.
Check it out: https://t.co/n9l8Jtho2Y
https://t.co/s3mPks9bgg

[// Lofty @loftyPuma]

Nobody else has a model so good that a world superpower called dibs

[Leon Brooks @leonbrooks79]

I’m a Codex users,

Why should I switch to Claude?

[// Lofty @loftyPuma]

@i_mika_el a condition

[// Lofty @loftyPuma]

@huge_icons As a famous Matt once said 👇

https://t.co/MfqxYh5Yvl

[Matt Pocock @mattpocockuk]

@DanielSmidstrup Your skill level acts as a ceiling for the agent

So every skill you gain multiplies your output

Never been a better time to git gud

[// Lofty @loftyPuma]

@araseb_ VCS that doesn’t fight your clankers

https://t.co/seCd6i9m7W

[// Lofty @loftyPuma]

@Arithmos0x @jcrew_eth @AIWayfinder @ParallelTCG @mookahfr You’re reducing Wayfinder to merely a bot?? Wayfinder is a ⚙️✨ 𝒸𝓁𝒶𝓃𝓀ℯ𝓇 ✨⚙️

[// Lofty @loftyPuma]

@Arithmos0x @jcrew_eth @AIWayfinder @ParallelTCG Can you make a path that plays parallel for me? I want to take in Prime and Prompt in one shot

[// Lofty @loftyPuma]

@antonships Yeah pretty much 😅 it improved significantly with fable so that was a fun 3 days. Gonna miss that

[// Lofty @loftyPuma]

All of the above haha
Opus plan quality is just better but codex works faster and better on most coding tasks. I even let opus act as the orchestrator and decide what effort level codex should use for each task.
I’ve had that setup running for a few weeks and I’m really happy with the results. I step in and guide the general code architecture here and there but it’s close to producing code I would write. I have a bunch of coding rules to push it that way but it’s all part of the process

[Rhys @RhysSullivan]

last one

[// Lofty @loftyPuma]

Fable is very smart. Smart enough that its output plans can raise the quality of other models just by handing off work.
So no, we can’t agree. Fable measurably produced better, cleaner code from other models (especially Opus and Sonnet, 5.5 needed some massaging for Fable to figure out how to prompt it but we got there after a couple tasks)

[Melvyn • Builder @melvynx]

Please be serious about Fable: for day to day coding, can we agree it doesn't change a lot and the hype is 10x the real value or no?

[// Lofty @loftyPuma]

@enjojoyy ngl it was really good. Not quite “have a world superpower call dibs” good but here we are

[// Lofty @loftyPuma]

Hey Dax! I’m building vcs for agents first. I think this is right up your alley, it is the lib first and harness agnostic.

Primary features for OpenCode:
- threads (worktrees but good)
- atomic undo/redo (agents can rewind to a given tool call or fork from a prior state or whatever)
- agentic attribution (know who delegated what to what agent)
- anti-drift knowledge (critical context embedded in the codebase and injected into the harness on read, it even uses semantic parsing to follow moved/renamed funcs)

https://t.co/n9l8Jtho2Y

[// Lofty @loftyPuma]

New life goal. Build something so good that a world superpower calls dibs.

[Anthropic @AnthropicAI]

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.

The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.

Access to all other Claude models is not affected.

We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.

Read our full statement: https://t.co/bwn0sximKZ

[Anthropic @AnthropicAI]

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.

The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.

Access to all other Claude models is not affected.

We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.

Read our full statement: https://t.co/bwn0sximKZ

[// Lofty @loftyPuma]

Well this aged like milk

[Alfie Carter @AlfieJCarter]

Fable 5 just changed outbound.

All you need is:

- A laptop
- A Claude Max subscription
- 10 minutes to install once

That's enough to run a complete outbound stack covering prospect, research, sequence, CRM, enrichment, deliverability, and pipeline review.

No expensive sales tools. No building prompts from scratch. No remembering which skill handles which task.

Most people still use Fable 5 for one-off prompts...

But the engineers pulling consistent pipeline are quietly using Fable 5 to:

- Build a prospect list from ICP filters in under 2 minutes
- Research any account and return a ready-to-use brief before every call
- Write a 5-touch sequence calibrated to the specific signal that triggered the outreach
- Run a pipeline health check and flag every deal with no next step, no recent activity, or a close date that has already passed
- Route any outbound task through one router skill that chains the right skills automatically without you remembering what anything is called

Usually, I sell this 89-skill bank for $97...

But today you can get it FREE.

Inside you'll discover:

89 skills filtered from the full sales-skills/sales repo for engineers doing outbound and pipeline work — everything not relevant has been cut

The /sales-do router skill that takes your objective in plain language, asks a few multiple-choice questions, and returns a detailed copy-paste prompt with your full context loaded

Platform skills for Salesloft, Apollo, Lemlist, Smartlead, and Instantly with prompts calibrated to each tool's native workflow

CRM skills for HubSpot, Salesforce, Attio, Pipedrive, and Close covering deal inspection, pipeline hygiene, and forecasting

Multi-step workflow chains that take you from a company name to a prospect list, enriched and verified, to a reviewed ready-to-send sequence in one session

Want it?

Like this post

Comment "SKILLS"

Follow me to receive it in DM

Available FREE for 48 hours only.

[// Lofty @loftyPuma]

Imagine how good GitHub would be if their incident reports looked like this

[Polar @polar_sh]

Our deepest apologies for the downtime today between 13:17 and 14:18 UTC. Tens of thousands of businesses trust Polar with their billing, and an incident like this should not happen. We'll publish a full post-mortem in the coming days and take actions to make sure it does not happen again. In the meantime, here's exactly what happened.

We rotate secrets frequently as a security policy. Today we rotated our database credentials – but a flaw in our config, where the credentials were tied to the database's Terraform resource, caused Terraform to trigger a full resource replacement. Our production database was replaced. It took us an hour to restore and bring everything back.

Was any data lost? No. We run point-in-time recovery with remote backups, hosted on Render with enterprise support. We immediately began a PIT recovery, replaying WAL up to the moment before the Terraform change applied. Recovery completed at 14:16 UTC; we confirmed full data integrity and brought infrastructure back online at 14:18 UTC.

A few edge cases we're reconciling now – none involve lost data:

— Our API stayed up for ~50s after the latest point in time recovery point and kept accepting event ingestion (which we double-write into Tinybird). A few hundred events landed there; we're going to replay them.
— We received 3 Stripe webhooks for in-flight transactions as we went down. We're resolving these and reaching out to the affected customers directly.
— After restore, our webhook worker was properly restarted to point to the recovered database. That's fixed; it's catching up now, so webhook delivery is delayed but recovering.

Polar is fully operational. The only residual effect was delayed webhook delivery while the worker catched up, which it has now. No action is required on your end.

Our post-mortem will detail the full timeline and the guardrails we're adding so this class of failure can't recur, and we’re truly sorry it happened in the first place.

[// Lofty @loftyPuma]

@DamienWayne VCS that works for your agents instead of against them
https://t.co/seCd6i9m7W