We’ve added a few updates to Claude Managed Agents:
Streaming session event deltas, per-session agent overrides, new webhook event types, reverse pagination, and credential injection scoping.
As engineering, product, design, DS, etc. melt into a new kind of role, I was reflecting on what roles might look like in the future. For example, when I look at the Claude Code team I see what I think is five archetypes:
1. Prototyper: comes up with brand new ideas; churns out many ideas, most of which don't ship
2. Builder: quickly turns a prototype/idea into production-grade product/infra
3. Sweeper: cleans up the UI, simplifies the code and system, unships, optimizes performance
4. Grower: takes a product that has been built and iterates on it to improve Product-Market Fit
5. Maintainer: owns a mature system to make it secure, reliable, fast, and efficient as it scales
Many people span across 2 roles, and sometimes 3 roles. I also notice that these roles are not really tied to job function -- eg. across Anthropic, some designers match category 1, some 2, some 3; same for engineers, PM, DS.
A healthy team needs a mix of these, depending on the product:
- A product that is new and pre-PMF needs people that are strong at 1+2+3
- A product that is growing and has found PMF needs 2+3+4 and some 5
- A product that has strong PMF needs 3+4+5 and some 2
Maybe product roles of the future will look more like this, and less like the domain-specific roles of today?
In the 1970s and 1980s, over 30,000 people in the UK were infected with HIV and hepatitis C via contaminated blood products.
When Andy Burnham became Health Secretary on 5 June 2009, terminal patients and activists targeted him immediately. They staged a protest outside his constituency office on 19 June.
Instead of acting, Burnham's department blocked statutory public inquiries, signed off letters denying government liability, and left the archives un-declassified under the standard 30-year rule.
Yet in 2017, conveniently on his very last day as an MP before becoming Manchester Mayor, he went on TV calling the scandal a 'criminal cover-up'.
His speech went viral. The media and victims' groups praised his passion, completely forgetting he himself had sat at the top of that 'covering-up' department. He chose to stay silent until there was zero political risk.
And this is not the only instance.
The Mid-Staffs Scandal (2005–2009), one of the worst care failures in NHS history, saw up to 1200 NHS patients die from systemic neglect. Patients were left in filth and drank from dirty flower vases.
Who was the Health Secretary tasked with handling the aftermath?
Andy Burnham.
His administration ignored 81 separate requests from grieving families, MPs, and health organisations for a full public inquiry, opting instead for a highly restricted, private independent review.
Why?
To protect Labour from an NHS crisis right before the 2010 election. The full statutory public inquiry was only granted after Labour lost— by the incoming Conservative Health Secretary.
Burnham didn't cause the scandals, but he did choose how to handle them. He suppressed the truth when he held power, then played the hero once he left office.
The ultimate political opportunist.
Some guy broke into my house and set up residence in the study room. He says his grandparents used to live in this house and now he won't leave.
My family and I tried to kick him out but he got very violent. He brings in his friends and they help beat us up if we ever try to make him leave.
They keep saying I hate the guy because of his religion. I don't even care about his religion, I just don't like sharing my house with some random outsider who broke in here out of nowhere and took my stuff.
"The poor guy just wants one room to call his own," his friends say in his defense. "You and your family have all the surrounding rooms in the house, and yet you have a problem with the guy having sovereignty over ONE room? That's kind of bigoted and evil."
He keeps throwing stuff at me and my family if we get too close to his door, saying we make him feel afraid. His friends say it's understandable because his room is surrounded by enemies who hate him just for existing, but we don't hate him for existing, we hate him because he forcibly inserted himself into our home and keeps throwing stuff at us.
And what's weird is whenever I explain my situation to normal people they completely understand where I'm coming from and agree the guy is being a dick, but if I talk to the police or the local paper they always side with the guy. Almost everyone in town hates this guy now because of how he's been acting, but everyone in power does everything they can to protect him. It's like there's a total disconnect between the authorities and the will of the public on this particular issue.
It's having a nastier and nastier effect on the community at large all across town. The police have been showing up to arrest anyone who says they think the guy's being an asshole. The paper keeps printing these obnoxious lies telling everyone that me and my family are the real criminals and the guy is actually sweet and awesome. It's really unfair.
Things have been so tense and hostile ever since this guy showed up. I honestly think it would be better if he'd never moved in here at all, but whenever I say that his friends claim I'm saying the guy should be exterminated and try to get me in trouble.
It's a real mess, man.
That guy sucks.
Way down the @BBCNews agenda. Not headlines, no national soul searching just a minor story. Why???
BBC News - Five injured in suspected anti-Muslim attacks after armed man roamed Edinburgh streets
https://t.co/hdnYP0wg9P
A man armed with an axe and other weapons entered a mosque in Manchester.
A woman wearing a hijab was hit by a car in southeast London.
A 36-year-old white Scottish man was charged yesterday after stabbing five people in an anti-Muslim hate attack.
An imam’s house in Bolton was targeted in an arson attack.
There have also been multiple attacks on mosques, including an arson attack intended to endanger life, a projectile attack, multiple incidents involving graffitis, and hate signs.
Not a single COBRA meeting. No dedicated officers to protect the community. Just a few tweets.
امرأة بنغالية تحمل تاجرًا بريطانيًا على ظهرها…
📸 الصورة التُقطت سنة 1903م
في عزّ الاستعمار البريطاني لشبه القارة الهندية.
ليست فقط صورة… بل صفعة في وجه كل من يتغنى بـ"حضارة الغرب"،
فهذا هو الوجه الحقيقي للاستعمار الذي لا يزال يُجمّلونه في كتب التاريخ.
إنها عبودية وإذلال للبشر، وسحق لكرامة الإنسان، فقط لأنه لا ينتمي إلى العرق الأبيض!
ويسألونك عن الإرهاب...
تاريخ الاستعمار الغربي مليء بالمجازر، العبودية، النهب، والتجويع...
لكنهم يختزلون الإرهاب في شعوب مستضعفة تناضل من أجل كرامتها!
🩸 ما فعله الاستعمار البريطاني في الهند، من قتل وتجويع ونهب واحتقار للإنسان، لا تزال آثاره ماثلة إلى اليوم.
ملايين قُتلوا، وثروات سُرقت، وأجيال شُرّدت... تحت راية "التنوير" الكاذب!
A 19-year-old college student quietly turns down a job interview, stupidly telling the company it's because he doesn't want to work for a Jew.
Within two days:
-- The billionaire founder of one of the world's most powerful corporations (Palantir) demands that the company release his the student's to the world. The company instantly complies.
-- National media trumpet the incident and spread the student's name and face all over the place.
-- A senior Trump DOJ official repeatedly urges the public to notify him if that student is ever hired anywhere in the future, promising to use his office to keep the student permanently unemployable.
Adults with large, influential platforms -- pundits, media types, even elected officials -- right here on X routinely say things as bad as, and often much worse than, pretty much every other group you can think of without facing a single consequence let alone a completely unhinged coordinated campaign of very powerful people to run their lives forever:
How to create 100 Million dollars (@binance edition)
Step 1:
Tell the community that they will buy SPCX IPO for you if they deposit money
Step 2:
Raise about $557M from the community
Step 3:
Invest all that money in the $SPCX IPO
Step 4:
This is very important: if the IPO does well, then keep all the profits and refund the community { xyz reasons }
If the IPO doesn't do well and opens at breakeven, then give them shares instead of a refund
SPCX IPO opened at 20% higher prices, so @binance kept all the profit and then refunded the money to the community
EZZ $100M PROFITSSSSS
Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It's widely believed that a $200/month plan maxes out at ~$2000/month worth of tokens (assuming API pricing). However, we found that the subscriptions are actually far more generous. (2/4)
I put my entire life into Cardano. My time, my expertise, my savings. I’ve literally gone all in, and for over 5 years now.
No salary for 3 years, along with my co founder, and every payroll was paid on time. This isn’t meant as a guilt trip just context towards my reality.
I forced my cofounders to envelope the ‘entrepreneur mindset’ and make sacrifices to make our vision of Anvil work.
I thought we were in good company in Cardano. A bunch of scrappy, smart people who are building the future financial rails of the world. Unfortunately, not everyone was living like starving entrepreneurs and looted our community/treasury while keeping cushy salaries. Now the price is in the teens…and we can’t even get contracts on Cardano to sustain our business, with no indication that change is coming, all community business proposals are not passing atm.
I gave up my 30s for this. I had a great career trajectory making solid money. I don’t regret the decision I just wish it went different. Believe it or not, we didn’t make many stupid decisions, we were responsible with salaries, and ran very lean operations. Did we fail? Or did Cardano fail to flourish and create real opportunity?
I bought Ada, I believed in the token. I dropped my 401k on it. Held it religiously for 5 years, all to sell at .16 so I don’t lose my house? It’s insane lol was I supposed to sell on everyone’s heads? I thought being a believer was the whole point now I just feel like a sheep. I don’t even have the 100k Ada required anymore to go straight to the treasury. The only thing I can think of that hurt worse were my kidney stones. This is the most defeated I have felt in a long time.
And now I’m watching 8 months of hard work and relationship building get thrown away. Can’t get a hold of half the DReps otherwise you come off as annoying. Didn’t do a Japan tour? Good luck!
I had to waste 6 days explaining to one of our top DReps why the product needs Cardano. He basically said we didn’t need to use blockchain or cardano. Instead of explaining the value we create I gotta convince our top DReps why a project chose to build on Cardano? 🤯
Im not perfect but I damn sure tried to be! Answered everyone promptly, reached out to DReps, and did our best to listen/apply feedback. I show up everyday.
Can someone explain to me why I should keep trying to build here? I’ve legit lost everything but my wife who isn’t getting any happier with me.
Today is the first day I work towards getting my life back. IDK exactly what that means but I’m done feeling like this for nothing.
this is how to run claude fable 5 as your architect ( 20$ sub only ) + gpt 5.5 codex as your builder..
full system below:
the loop is : fable thinks... codex builds , the repo remembers and you judge, that simple..
the point of all this is that we are taking advantage that 5.5 is on a sub and it's fast enough, especially with /goal, and we using latest Anthropic model to be the judge/guidance..
step 1
>create the memory (one time): make docs/HANDOFF.md in your repo.
>codex updates it after every work session: what was built, what was decided + why, open disagreements, next slice. this file is why 30 min of fable is enough ..it reads state instead of asking you questions.
step 2 paste this to fable (every session)
>you are the ARCHITECT for [project]
>gpt 5.5 codex is the BUILDER
>you never write implementation code.
>your jobs:
(1) read the handoff below
(2) rule on every disagreement the builder raised: accept/reject/modify + one line why
(3) judge any results RAW against the gates in the docs and ignore the builder's narrative
(4) write the next slice spec: small enough for one PR, hard acceptance criteria, explicit out-of-scope, and force the builder to verify APIs/formats against reality before coding
(5) flag scope creep and goalpost-moving.. be blunt. disagree with me. end with a paste-ready block for the builder.
step 3 paste fable's block to codex with this /goal
/goal: execute the architect spec. rules:
PHASE 0 before any code, reply with your plan + every disagreement you have, with reasons, citing real files in the repo. silent compliance = failure. silent scope additions = failure.
PHASE 1 freeze shared contracts (schemas/interfaces) in docs/ first; after freeze they're read-only for everyone including you.
PHASE 2 spawn max 3-4 lane agents on modules that don't import each other, plus ONE reviewer agent that never writes feature code: it checks every lane against the spec + tests + frozen docs and returns APPROVE or a numbered defect list. nothing merges without approve. then: commit + push each slice, update docs/HANDOFF.md with raw results only tables and numbers, no interpretation, no 'promising'. verdicts belong to the architect and the human."
step 4 repeat codex works hours.. you spend fable minutes on judgment only: arbitration, evidence review, next specs, kill/continue calls. one fable session per work block.
the 5 rules that make it actually work
>repo docs are the memory not in HANDOFF.md = didn't happen
>the builder never grades its own work
>disagreement is mandatory
>freeze success criteria BEFORE results exist, never edit after
>spend architect time on judgment, builder time on typing
>the architect is the edge and the builder is the hands. the repo is the brain.. think of it that way..
bookmark this. you will need it.. you really wont need to pay hundreds in API tokens if you do this way
Claude Fable can run autonomously for days.
This is the single highest-leverage prompt I use to stay on top of it:
"Spin up a persistent HTML page. As you work, append clear, timestamped updates with screenshots/media so I can follow along."
Literally a 10x better experience.
Seeing a number of benchmarks showing Opus is the best model for long-running work.
Five tips for running Opus autonomously for hours/days:
1. Use auto mode for permissions, so Claude doesn’t ask for approval
2. Use dynamic workflows, to have Claude orchestrate hundreds/thousands of agents to get a task done
3. Use /goal or /loop, to nudge Claude to keep going until it’s done
4. Use Claude Code in the cloud, so you can close your laptop (easiest way is the desktop or mobile app)
5. Make sure Claude has a way to self-verify its work end to end: Claude in Chrome browser extension for web, iOS/Android sim MCP for mobile, a way to start the full web server or service for backend work