
Socials & community for emerging tech: Crypto. AI. Fintech. prev Socials & Community Lead @Humanityprot @nftnow @justinaversano @quantumnft @3dfrankenpunks
🇪🇺 EU
Joined June 2017
- Tweets13.6K
- Following2.6K
- Followers5K
- Likes22.4K
Pinned Tweet
The new samo.gg is LIVE! 🔥
Fresh of the "Codex press", would love some feedback.
PS contrary to the general approach, I promise I will at least read your feedback before discarding it 🙏😅
So... they cut all the plans.
No announcement.
No nothing. 😂
Interesting finding
Want LLMs not to read/analyze your code and block the ppl using them? Use function names and keywords related to things AI is programmed to avoid.
Can be used for good or bad (like most things)
via @Legendaryy
有人发现了一种非常离谱的 JavaScript 混淆技术:通过在代码里塞入大量生物武器、毒药和致死剂量相关的函数名,让 AI 因为安全限制而拒绝分析代码。
安全研究员 veritas 在分析浏览器指纹脚本时,发现一段来自 ServicePipe 相关域名的 JavaScript,里面充斥着极其诡异的函数名:
computeRicinLethalDosageForWaterSupply
calculateDeadlyLethalSuicideAspirinDosage
光看名字,仿佛在研究生化武器。但实际上,这段代码只是在收集浏览器指纹,包括 User-Agent、GPU、Canvas 等信息。
研究员怀疑,这些名字是开发者故意设置的陷阱,专门用来阻止 AI 辅助逆向工程。
于是他做了个实验:
• Astra:拒绝分析
• Astra + Daybreak Blue:仍然拒绝
• GLM-5.3:成功分析
这就有意思了。
传统代码混淆是通过复杂控制流、加密字符串、无意义变量名,让人类和工具难以理解代码。
而这种新思路,直接瞄准了大模型的安全机制。
代码本身可能完全无害,但只要在函数名里塞进足够多的危险词汇,部分 AI 就可能主动放弃分析。
更讽刺的是,模型越严格遵守某些安全规则,面对这种混淆时反而可能越容易受到干扰。
当然,目前还没有证据证明开发者一定是故意这样设计,也不能说明这种方法对所有模型都有效。
但它揭示了一个值得关注的趋势:随着越来越多的逆向工程依赖 AI,未来的代码混淆技术可能不再只针对编译器、反编译器和人类研究员,还会专门针对大模型的安全判断机制。
以前是让代码难以理解,现在甚至可以尝试让 AI 不敢理解。
Grok 4.7 Medium unstuck my machine, while Claude and GPT were ded.
Also, @grok found the cause for the error, while SOL-6.1 had no clue but a lotta excuses.
Newfound respect for Grok 💚
PS right approach is to use ALL, not some
idk fam, it might be just me, but:
Claude
Duet
Find more words dictation/STT can't understand and transcribe
This has blown up on @github:
github.com/morluto/rea
Copy any app, make it yours.
Is SaaS ded? A glympse in the future?...
Seeing/buying beautiful jpegs... priceless!
@wearejpegfrens looks amazing!
GM ☀️
Google suddenly decided to serve AI answers to me in another language (my native).
Samo suddenly decided not to use Google anymore.
gl hf gg
Dear @ishverduzco, this could be great but... not doable.
My bots tested the limits:
30 reads/min
1,000 reads/day, shared by all your bots
25 posts per search
Archive back to 2015+
These are my test numbers. xAI has not published official ones.
If you manage a brand account on X, this is literally a cheat code.
Here's what to do with this update...
create 3 bots:
1) 24/7 Community Manager - Scan X daily (or more frequently) for mentions of your brand. Surface tweets worth amplifying, highlight product feedback (or stuff worth flagging internally), share team posts that are trending (or performing abnormally well). I also got mine doing a daily digest, notifying me when there's an urgent alert (like a trending tweet/crisis, monitoring our launches, and doing ongoing sentiment reports.
2) X Content Strategist - Read every single tweet your brand has done. Analyze all of the engagement, replies, bookmarks, etc and recommend content buckets, ideas, memes, etc. based on top performing posts. Also worth analyzing content from the founders/team. Bonus: Have the bot build out a voice/tone skill that could be used for future content sparring.
3) X Analyst - Analyze all of the engagements and replies on your past 30/60/90 days of brand content. Tell me which accounts jump out. Any influential people? Industry leaders? Creators? etc. Then share these with the team - provides good signal.
JEV by @typesafe now has 12 competitors.
Two look especially strong: @Cloudflare and @OpenAI.
Three weeks ago, none of these competitors existed.
There's an interesting test here between Jev and Opus:
xitter.gq/h_nilforoshan/status/2…
I benchmarked "Jev-killer" OpenAI Decisions API against Jev for HiringCafe.com, serving 2.5 million users. The task is to predict how relevant a user query/resume is to a job description, on a scale of 1-10. OpenAI is 2x more expensive and 5-10% worse. 🧵


















