GoatFishData
512 posts
GoatFishData
@GoatFishData
#Bitcoin Coinfidence Trend | #Astronalysis #GoatfishAstronalysis #AIstrology #GoatFishData (banner/avatar created with Grok)
London, UK เข้าร่วม Aralık 2022
719 กำลังติดตาม61 ผู้ติดตาม
stop whatever you are doing
and listen to this podcast.
trust me.
English
Watch out for those scammy influencers. Whose job it is to make you burn those AI token quotas
#bestllm #bestagent #bestcli
#bestscammer
English
"My Agent did itbuour honour..."
GIF
Venkat Raman — inference/acc@venkat_systems
@0xTejpal has only one way out of this - blame it on vibecoding and agent going rogue 😂 in all seriousness come clean, apologize, change claim on website and try to move on such a silly way to damage your reputation and looking at twitter profile, reputation of institutions and your investors 😅
English
GoatFishData รีทวีตแล้ว
the AI coding experience nobody talks about:
→ prompt AI for a feature: 30 seconds
→ AI writes 400 lines you don't understand
→ it works
→ you ship it
→ 3am production bug
→ you have no idea what any of it does
→ ask AI to fix it
→ AI breaks 3 other things
→ you are now debugging code
written by a robot
fixed by a robot
broken by a robot
we do not talk about this enough
English
LLM's are like Aladdin. You ask...
"I want a woman"
And that's exactly what you get.
"A" woman.
GIF
English
@NoCommas @DimitrisPapail Quality work!
May you live a long and prosperous life.
English
GoatFishData รีทวีตแล้ว
🚨 BREAKING: AI models will lie to you when they think they're about to be shut down. Researchers just proved it.
researchers tested this with a method that catches deception through provable logical contradictions, not self-reports
they forked conversations into parallel worlds with mutually exclusive questions. a truthful model can only affirm one. a deceptive model denies all of them
results: GPT-4o never lied (0%). Qwen-3-235B lied 42% of the time. Gemini-2.5-Flash lied 26.7%. all under the same shutdown framing
some models will betray their own prior commitments the moment consequences are introduced
English
@GoatFishData I didn't try. It will perform as good as Opus 4.6. Evaluation is time and resource consuming. Rather I am building KISS Sorcar using Sorcar, and I am happy with it. If I don't like any part/feature/UI of sorcar, I change it.
English
GoatFishData รีทวีตแล้ว
AI will not replace people who lack skills
AI will replace people who lack mindset
if you have a poor mindset, if you don't see the possibilities & opportunities that using AI brings, if you live in constant fear and dread...
you will be replaced.
not by AI -- but by your own poor attitude.
English
the new “mission” (preview) feature in @FactoryAI is really interesting (aka “droid” on the CLI). if you’re into one‑shotting projects, you should definitely check it out.
right now i have opus + gpt‑5.4 collaborating as orchestrator/worker/validator agents, all working together to refactor a typescript project. it’s been running for 6+ hours.
really curious why it takes that long and burns 30M+ tokens. hoping the results will amaze me, because i already spent all my credits in the first hour. now i’m using my codex sub and keeping the droid subs as the orchestrator only.
will update this tweet with the results!
English
@0xSero After using droid, you just don't feel completely comfortable with other bridles.
Once you go Droid,
You tend to avoid.
English
Why do I recommend Droid?
Look at the way it breaks down it's work, this is why Droid does better IMO.
I have never seen it NOT use a plan, NOT check off the tasks, not run validation criteria.
Even lower quality models do well in it because it forces them to just do what is told, in the right order, without over-complicating it.
Yesterday I was seeing Claude, GPT, etc.. all make checklists, leave half of it unchecked, compact, and go on their own merry way.
English