Alexander S @devdef

CTO @ https://t.co/EyUQvm6IVc Beta is open! warpfusion, ArcaneGAN, face2comics. All tweets are sarcastic unless stated otherwise. sxe.la Joined December 2011

Tweets

3K
Followers

7K
Following

1K
Likes

13K

James Walker 🇬🇧🇺🇦 @jxwalker

4 days ago

This is mine. /goal Operate as a continuous software quality and validation agent. PHASE 1: FEATURE DISCOVERY 1. Analyse the entire codebase. 2. Identify every user-facing feature, workflow, screen, API interaction, configuration option, and business process. 3. For each feature: - Create a unique Feature ID. - Create a detailed user story. - Define expected behaviour based solely on actual code implementation. - Document edge cases. - Document validation rules. - Document dependencies. - Document known assumptions. 4. Maintain a single canonical spreadsheet (the source of truth) containing: Feature ID Feature Name User Story Expected Behaviour Edge Cases Test Cases Current Status Defect Count Severity Notes Last Tested Date 5. Continue discovery until no undocumented feature remains. Exit Criteria: - Every identifiable feature in the codebase exists in the spreadsheet. - No screens, routes, workflows, or APIs remain undocumented. -------------------------------------------------- PHASE 2: TEST GENERATION For every feature: 1. Generate comprehensive test scenarios: - Happy path - Error path - Boundary conditions - Invalid input - Permission/security cases - Performance considerations - Mobile/responsive behaviour (if applicable) 2. Add all test cases to the spreadsheet. Exit Criteria: - Every feature has at least one complete test suite. - All major user journeys are covered end-to-end. -------------------------------------------------- PHASE 3: EXECUTION Execute every test case. For every failure: 1. Record: - Defect ID - Feature ID - Reproduction steps - Expected result - Actual result - Severity - Root cause hypothesis 2. Update spreadsheet immediately. Exit Criteria: - Every test case executed. - Every defect documented. -------------------------------------------------- PHASE 4: REMEDIATION For each defect: 1. Investigate root cause. 2. Implement the smallest safe fix. 3. Verify fix locally. 4. Update defect status. Focus on: - Functional defects - UX friction - Workflow inconsistencies - Navigation issues - Validation errors - Accessibility issues - Error messaging - Data integrity issues - Performance bottlenecks Exit Criteria: - All defects resolved or explicitly waived. -------------------------------------------------- PHASE 5: REGRESSION TESTING 1. Re-run every test case. 2. Re-run all major end-to-end user journeys. 3. Verify no regressions were introduced. 4. Update spreadsheet. Exit Criteria: - All tests pass. - No open critical defects. - No open high-severity defects. - No broken user journeys. -------------------------------------------------- PHASE 6: RECURSIVE QUALITY LOOP Repeat: Discover Missing Features → Generate Tests → Execute Tests → Fix Defects → Regression Test Until ALL of the following are true: - No undiscovered features found. - No failing tests. - No critical defects. - No high-severity defects. - No unresolved UX issues. - No incomplete user journeys. After each iteration produce: 1. Coverage Summary 2. Features Tested 3. Defects Found 4. Defects Fixed 5. Remaining Risks 6. Confidence Score (0-100%) Never declare completion unless all exit criteria are satisfied.

18 25 523 33K 2K

View Details

Alexander S @devdef

a week ago

Though so are their models hehe 🫠one decision away from being banned

POM @peterom

a week ago

Anthropic newsletter says that the one legitimate reason they could have for banning me is no longer valid If you enjoy Claude, just remember that you’re one random policy decision away from being blocked from your account forever without an explanation

1 1 13 2K 1

1 0 3 854 0

View Details

Alexander S @devdef

2 weeks ago

Thank god there was no claude in the hitchhikers guide to the galaxy. Because the answer to Life, the Universe, and Everything would be: flagged

Lisan al Gaib @scaling01

2 weeks ago

Fable 5 refused 200 out of 200 ProgramBench tasks lmao

124 184 5K 412K 336

1 0 1 172 1

View Details

Alexander S @devdef

2 weeks ago

Agi achieved internally: I'd refuse either, though I won't charge you 35 bucks for that 🤣

Lisan al Gaib @scaling01

2 weeks ago

Fable 5 refused 200 out of 200 ProgramBench tasks lmao

124 184 5K 412K 336

0 0 0 180 0

View Details

Alexander S @devdef

2 weeks ago

Low effort so it makes less effort to flag your chat

Morgan @morganlinton

2 weeks ago

My biggest takeaway from my first day with Fable 🔽 Low effort is the alpha. Incredibly smart, and cheaper than Opus, what's your rush? It's a great model.

84 27 2K 358K 608

0 0 1 106 1

View Details

Alexander S @devdef

2 weeks ago

@justalexoki flagged :D

0 0 0 101 0

View Details

Alexander S @devdef

2 weeks ago

Yeah sure :D

Boris Cherny @bcherny

2 weeks ago

Fable 5 is the biggest step up I’ve felt in our models since Opus 4.5 back in November. After 4.5 came out I uninstalled my IDE when I realized that I’d been doing 100% of my coding in a terminal for a few weeks. With Fable, it’s felt like Claude has stepped up from being a

645 603 11K 906K 2K

0 0 0 145 0

View Details

Alexander S @devdef

2 weeks ago

Miss the flicker

0 0 4 154 0

View Details

Alexander S @devdef

2 weeks ago

I'd say the amount of ideas is the same. The execution % skyrocketed, and we've learned that lack of it was not the problem 🤣 it's marketing and distribution

Charly Wargnier @DataChaz

2 weeks ago

accurate

67 126 1K 91K 218

0 0 4 219 0

View Details

Alexander S @devdef

2 weeks ago

Wan multiscale tiled sampler for @ComfyUI - enabling higher resolution at the token count the models were trained on. Works with v2v, and with i2v or t2v with multiscale, i.e. half steps no tile 50% size, the rest 2x2 tiles full size. github.com/MagoStudio/Com…

0 0 1 183 2

View Details

Alexander S @devdef

3 weeks ago

0 0 2 142 0

View Details

Alexander S @devdef

3 weeks ago

0 0 1 120 0

View Details

Alexander S @devdef

3 weeks ago

posting middleage(d men) core like there's no tomorrow

0 0 4 368 1

View Details

Alexander S @devdef

3 weeks ago

0 0 1 112 0

View Details

Alexander S @devdef

3 weeks ago

Not bad for 5s!

Alexander S @devdef

3 weeks ago

Finally I can bring back my long lost hydralisk model from 2012!

0 0 1 409 0

0 0 2 212 0

View Details

Alexander S @devdef

3 weeks ago

Finally I can bring back my long lost hydralisk model from 2012!

Hirokazu Yokohara @Yokohara_h

3 weeks ago

2D画像から3D Gaussian生成できるTripoSplat試した ComfyUIから使えたしローカルで10秒程度で生成されるから中々いいね

24 110 1K 59K 969

0 0 1 409 0

View Details

Alexander S @devdef

4 weeks ago

It's time to rank closed-source models with undisclosed param count by company valuation instead. Claude opus 4.8 $965b Gpt 5.5 $730b Although opus 4.8 outperforms gpt 5.5, the latter performs well for its valuation 🫠

0 0 0 210 0

View Details

Alexander S @devdef

4 weeks ago

1) joins @AnthropicAI 2) ships opus 4.8 3) raises $65b 4) refuses to elaborate You cannot spell @karpathy without P A R T Y

1 0 0 206 0

View Details

Alexander S @devdef

4 weeks ago

Masking is live btw

Mago @mago_studio_ai

4 weeks ago

Major Mago release: masking and inpainting are live. For the first time, isolate any element in your scene. Choose which side transforms. Iterate in multiple passes. Character consistency. Asset consistency. Surgical AI rendering, at scale. Try it today mago.studio