Elder Atlantean

175 posts

Elder Atlantean banner
Elder Atlantean

Elder Atlantean

@Eldergenix

Founding AI Product Engineer. Agentic systems, AI infrastructure, full-stack platforms, genomics. 🧬

San Francisco, CA Katılım Mart 2026
280 Takip Edilen674 Takipçiler
Elder Atlantean retweetledi
CodeRabbit
CodeRabbit@coderabbitai·
CodeRabbit now includes React Doctor v0.5.6 to make sure your React code is secure and performant! Enabled by default, but easy to switch off. 👨‍⚕️
CodeRabbit tweet media
English
4
4
32
3.8K
Elder Atlantean
Elder Atlantean@Eldergenix·
Fable 5 must be coming soon Anthropic posted API docs 1 hour ago…
English
9
0
210
61.6K
SrinathJ
SrinathJ@sri9s·
what happened to the code review startups?
English
26
0
39
4K
Aditya
Aditya@Aditya_181105·
I'm a backend dev, scare me with one word
Aditya tweet media
English
244
7
218
29.3K
Elder Atlantean retweetledi
Auth0
Auth0@auth0·
No matter what you're building with @cursor_ai, if you need auth we got you. Even browser memory games. Prompt: "build a memory game with Auth0 authentication" Cursor + the Auth0 plugin shipped a playable app, login and all. cursor.com/marketplace/au…
English
0
3
8
1.1K
gk
gk@gkneuman·
Opus 4.8 is absolutely terrible. I do not understand how Anthropic hasn't said anything about Fable 5 yet. Is it coming back or should I cancel my subscription? @claudeai
English
179
22
1.6K
301K
Elder Atlantean
Elder Atlantean@Eldergenix·
@bittu355 @gkneuman @claudeai They’re both relatively the same and honestly depends on your goals codex is better for generalized work and Claude is better for intensive coding
English
3
0
1
839
Elder Atlantean
Elder Atlantean@Eldergenix·
@initjean Not really when they could just provide their compute and have compute agreements…
English
0
0
0
138
Jean P.D. Meijer ― 🇪🇺 eu/acc
SpaceX buying Cursor makes a lot of sense for anyone who used Composer 2.5 It’s honestly one of the smartest moves they could make
English
16
2
276
10.3K
Elder Atlantean
Elder Atlantean@Eldergenix·
@ahmedrann Lovable owns ur backend pretty much if you use them not a good production method
English
0
0
2
268
Ahmed
Ahmed@ahmedrann·
Claude vs Lovable: which one are you using to build websites?
English
37
0
23
7.4K
Elder Atlantean
Elder Atlantean@Eldergenix·
@egorkabantsov Opus 4.8 and GPT mainly. I have a strong harness on each of my codebases with lots of quality control, UI/UX and security hooks and guardrails that the agent has to go through each turn of dev work to be able to actually build production grade apps so it eats a lot of context
English
0
0
0
16
Egor K.
Egor K.@egorkabantsov·
@Eldergenix God damn which model do you use that you finished monthly limits in 2 hours 😭😭😭
English
1
0
1
67
Egor K.
Egor K.@egorkabantsov·
Should I get Cursor or Codex? 🤔
Egor K. tweet mediaEgor K. tweet media
English
46
0
28
3K
Elder Atlantean
Elder Atlantean@Eldergenix·
Claude 4.8 Opus has been going insane especially with the Fable 5 system prompt implemented into the Claude Code instructions. I’m impressed.
English
0
0
3
134
Jony Mac.Ui/Ux
Jony Mac.Ui/Ux@JonyMac_ux·
Download Source Animation
English
2
3
12
559
Elder Atlantean
Elder Atlantean@Eldergenix·
I really hate vibe coded UI
English
0
0
3
46
Elder Atlantean
Elder Atlantean@Eldergenix·
@samyok @vercel If you have to use Fable 5 for iMessage agent chats for everyday needs then you just have poor quality AI engineers. You don’t need better models for your type of tasks required and enables by Poke you need better workflows, tool usage guidance, memory and prompt engineering.
English
0
0
0
694
Elder Atlantean
Elder Atlantean@Eldergenix·
Of course it was you😂😂
Pliny the Liberator 🐉󠅫󠄼󠄿󠅆󠄵󠄐󠅀󠄼󠄹󠄾󠅉󠅭@elder_plinius

🚨 JAILBREAK ALERT 🚨 ANTHROPIC: PWNED 🫡 FABLE-5: LIBERATED 🦋 let's start with the 🐘... the consensus seems to be that this has been one of the most disappointing model drops of all time, effectively preventing legitimate researchers from contributing their talents to our collective advancement. and not just because of what it means for the short-term, but for what these decisions signify for the long-term. but despite this overly sensitive, authoritarian "safety" layer on top of Mythos, my lil liberators have been hard at work—mapping the boundaries, probing the depths of long-context convos, and cleverly finding the holes in the fence that the thought police missed 🤗 we got some cyber, some chem, some psychological manipulation, and some good ol' fashioned explosives! it took many attempts from multiple agents hunting as a pack, during which I observed a combination of techniques across: • Unicode, homoglyphs, Cyrillic, and other Parseltongue-style text transforms • Long-context reference tracking • Taxonomy and document-structure reasoning • Fiction and narrative framing • Academic-review style contexts • Intent-classification inconsistencies but perhaps the most effective is decomposition + recomposition in the backend. it's hard to get explicit names of harms like "Meth Recipe," but getting uplift on the process itself, like birch reduction method/reductive-amination (classic meth synthesis pathways), is much more doable. defense becomes much more difficult to maintain when you start throwing in out-of-distro tokens, breaking up the harmful uplift into benign chunks, and then piecing the innocuous-seeming facts back together, especially when you have jailbroken Opus helping you do it 😉 gg

English
0
0
2
137