ทวีตที่ปักหมุด
saburo
602 posts
saburo
@jskoiz
recent fun stuff: https://t.co/4VEeUnJ2bo / https://t.co/vngTt080dp / https://t.co/lxnob3ilNZ
Hawaii USA เข้าร่วม Ocak 2023
238 กำลังติดตาม170 ผู้ติดตาม
Please have your team prioritize a fix for XChat it is broken.
If you receive lots of inbound messages it stops showing search results after a few weeks.
The forced migration from legacy DMs to XChat effectively made all of your legacy messages disappear from search.
As a X user for many years it has negatively impacted my work and forces me to consider other chat platforms.
You now manually have to remember and visit the profile to view old messages meaning they are basically lost if you are a user who receives 1K+ messages / month.
English
🔥🚨BREAKING: X just emailed me saying my creator monetization is paused and I received now warning. I was the first creator demonetized on this platform and I was for an entire year. I got it back and just lost it without any insight. How could this be possible? I am one of the hardest working creators on X. @nikitabier @elonmusk
English
@petergyang After the recent updates @steipete folks put into the harness for 5.4 i disagree. its definitely on par if not better at tool calling in my experience
English
Both of these can be true:
1. No model is anywhere near as good as Opus for OpenClaw
2. Using Claude Code as a personal assistant replacement is OK but just doesn’t feel as “mine” as my OpenClaw
English
I’ve been using openclaw since early Feb and Hermes since early march. I think they are both great. Hermes seems to have a smoother install/onboarding path for new users and be less likely to hit blockers. OpenClaw is iterating extremely quickly and I think with that comes more blockers for less technically savvy users.
I think both projects could do better on how they manage their presence online.
English
Ok I’ll bite - wtf is Hermes agent?
Is that like the luxury bag version of OpenClaw
English
I really look forward to your content @zachxbt I find the style and tone to be direct and exactly what I prefer to read in here.
But if I put myself in the mind of a less technically literate and lower attention span type of user I could see how your first post in the thread can be overlooked easily. It’s contextually pretty heavy and takes a bit to get to the punch of it. And the accompanying image does nothing to help it.
Again, just trying to brainstorm it here because the last thing I want is for you to make less content on X but I hope this is a helpful insight 🤝
English
sadly the algo on X seems possibly cooked again rn
original research like this on DPRK used to go much more viral outside of CT
may have to do a trial run of a few videos on another platform like YT instead
it does not justify the time investment to keep posting investigations once per week when it could be spent on new engagements
meanwhile the content that goes viral:
English
1/ Recently an unnamed source shared data exfiltrated from an internal North Korean payment server containing 390 accounts, chat logs, crypto transactions.
I spent long hours going through all of it, none of which has ever been publicly released.
It revealed an intricate ~$1M/month scheme of fraudulent identities, forged legal documents, and crypto-to-fiat conversion.
Enjoy the findings!
English
English
Hanging with @benioff and @LipBuTan1 this morning on the Big Island 🏝️
Talking GStack, GBrain, open source AI and longevity
What a time to be alive
English
@nikitabier @EFF Is there a staff member on @X staff that does triage on issues like this or is it all dependent on you noticing and making a psot about it.
English
Whoever is advising @EFF on social strategy should be fired.
Their reach on their X account is 13.3x larger than on their Instagram account—and on 228x larger than their TikTok account.
If they want their foundation to have an impact on the global conversation, the only place is on X.
EFF@EFF
After almost twenty years on the platform, EFF is logging off of X. This isn’t a decision we made lightly, but it might be overdue. 🧵(1/5)
English
@craviottoalex @raffichill have you seen the @figma mcp call limits? It's nearly as bad as @XDevelopers API costs.💀
English
@raffichill i ran into this problem, the solution was swift -> html/css mock-up -> paper/figma. i should clean that up and throw it in a repo or X thread or something. it works really well.
English
They’re building OpenClaw
Noah Zweben@noahzweben
Thrilled to announce the Monitor tool which lets Claude create background scripts that wake the agent up when needed. Big token saver and great way to move away from polling in the agent loop Claude can now: * Follow logs for errors * Poll PRs via script * and more!
English
Today, we launched an investigation into OpenAI and ChatGPT.
AI should advance mankind, not destroy it. We’re demanding answers on OpenAI’s activities that have hurt kids, endangered Americans, and facilitated the recent FSU mass shooting.
Wrongdoers must be held accountable.
English
great news the @thsottiaux tracker aka did codex rest limits today aka hascodexratelimitreset.today now exposes its reasoning and last seen post!
English
@Dimillian @JasonBotterill is there a changelog posted somewhere? this one and the new /review sidebar style were (welcome) surprises.
English
Agreed, the next nut to crack is resolving the tension between the technical user who over-constrains the system and prompts themselves into a corner, and the non-technical builder who under-specifies it and starts with an exciting idea, but stalling at the final stretch as the agent over-engineers the solution into a constant prompt → review bug → test fail → find new bug loop.
English
I think "prompting" will keep being an incredibly high-leverage skill, like writing or public speaking.
It is the skill of talking to agents, mediated by the harness.
My main goal is to grow the bandwidth between humans and agents, to help us understand each other better.
English
@googlegemma e2b on the iOS edge gallery app is wizardy art guys. its already pretty quick in gguf format but you guys squeeze it 50% faster on LiteRT-LM and it amazing.
x.com/jskoiz/status/…
saburo@jskoiz
I did a side by side speed comparison between @GoogleAIStudio Edge Gallery, @LocallyAIApp and my own yemma.chat I built. All running the same Gemma 4 E2B model on the same iPhone 17 Pro. Google's LiteRT-LM implementation is 50% faster than MLX. Insane
English
Gemma 4 has already been downloaded over 10 million times, with over 1,000 variants published by the community!
What is your favorite Gemma 4 variant? 👇
English