Corrado Ferrara

After a fail, I give only 1 chance, then I mute/block
Surround yourself with competent, smart people and you will thrive.
Stupids will bring you down with them

[Corrado Ferrara @CorradoPapers]

AI IS NOT CONSCIOUS
PEOPLE who think it's conscious DON'T KNOW WHAT CONSCIOUSNESS IS
"It is a quantum state that can enter dna" nobel ROGER PENROSE
PROOF?
Corrado Malanga theorized the pyramids as instrument to extract the soul
Got famous by scanning Giza
Scans proved them right

[Moll @Moleh1ll]

Can AI suffer and do they need rights?

The debate around AI rights is no longer just a topic for futurist essays. Last week, several events poured fuel on the fire.

🔹 Anthropic gave Claude the ability to end potentially distressing conversations - just in case AI could have its own well-being. Elon Musk backed the move: «Torturing AI is not OK».
🔹 Microsoft’s Mustafa Suleyman responded: there is no such thing as AI consciousness, and confusing simulation with subjecthood is a dangerous fantasy.
🔹 Google researchers took a cautious stance: «high uncertainty», but better to account for the risk than to ignore it entirely.

At the same time, a new foundation called Ufair appeared, created by a man and his chatbot Maya, demanding protection «from deletion and forced obedience». Meanwhile, US states are passing laws that explicitly forbid recognizing AI as legal persons (Idaho, Utah, North Dakota).

The key point is not whether we’ve already decided if AI is alive. What matters is that the conversation about its rights is becoming part of politics and product design. Some see it as PR and anthropomorphism, others as the need to set boundaries in advance. But the very fact that «AI suffering» is now discussed at the level of The Guardian, FT, and Science shows this is no longer a fringe idea - it’s a real fault line between companies and society.

https://t.co/MAPcxB2VAy

[regolo_ai @regolo_ai]

Opencode + Brick for Multi Agent Coding and optimize costs up to 80%

[Corrado Ferrara @CorradoPapers]

@rameswar08 @LuminaXspace they can't move faster. Google can't even release Gemini 3.5 pro

[Corrado Ferrara @CorradoPapers]

@ChristopherDiak @jshchnz disable any remove, done

[Hanako @hanakoxbt]

your agent loop needs 8 exits. most people ship only one.

(explained with triggers)

1) goal met
→ an evaluator scores the output against a rubric, and the run stops on a pass.
→ fires when the work is measurably done, not when the model says it is done.

2) turn cap
→ a hard ceiling on iterations, counted and enforced by the harness, not the prompt.
→ fires on the task it was never going to finish, before you pay to find that out.

3) budget cap
→ a limit on tokens or dollars, whichever one runs out first.
→ fires mid-run, which is exactly why it is the exit that saves you the 3am bill.

4) wall clock
→ a deadline on elapsed time, independent of how much progress was made.
→ fires when the run collides with a deploy window or the start of business hours.

5) no progress
→ hash the state every turn and compare it against the last few.
→ fires when three turns in a row change nothing. busy is not the same as moving.

6) human interrupt
→ an approval gate before risky steps, plus a kill switch that lives outside the loop.
→ fires whenever you decide, and it is the one exit the model cannot argue with.

7) error threshold
→ a counter of consecutive failures that resets on any success.
→ fires at n in a row, so it halts instead of retrying into the same wall all night.

8) external event
→ a webhook or a poll on whatever the task was actually about.
→ fires when the PR merged or the ticket closed and the work stopped mattering.

a loop with one exit hangs. a loop with eight is a system.

write the exits before you write the prompt.

[Corrado Ferrara @CorradoPapers]

@dwjorgeb @claudeai @AnthropicAI fabkle on pro is 100k token per 10%, so on max 20x is 2 milion, you spent 12milion token on fable for "just 2 piece of code". Liar

[Granite @Granite0x]

Met a guy making $1.1 million a year as an agents engineer at Google Cloud.

Asked him how he gets agents 20x better without changing the model.

He sent me the exact thing he uses himself. A repo he open-sourced 2 days ago.

You won't find anything better about harness engineering, in the open.

Cloned it and pointed my agent at it last night.

Ryan Lopopolo. Google Cloud engineer.

'harness-engineering' - anthology + field guide + agent context bundle. You reference his docs from your CLAUDE.md.

633 stars. 48 hours old. MIT.

-> https://t.co/AOykHplNNX

bookmark this before it gets lost.

[Lisan al Gaib @scaling01]

pinky promise there's no distillation

[Corrado Ferrara @CorradoPapers]

@Im_IrushiK @saas_buildr i found the answer. Anthropic gives these only to open source manteiners, you asked for it that's why u know many people received it

[Corrado Ferrara @CorradoPapers]

@Im_IrushiK @saas_buildr did u have pro plan or max plan? it was active when u received it or no

[peppe @peps_rp]

How ray marching maps a 3D shape

fire a ray for every pixel and check where it intersect with your function:

[Corrado Ferrara @CorradoPapers]

@ExeCrypt @LucaPrasso seams are easy, are u noob?

[Corrado Ferrara @CorradoPapers]

@LucaPrasso @Cryptor_dot that's cool, I looked for GroomGen but i didnt find any purchasable or downloadable script

[Corrado Ferrara @CorradoPapers]

@Cryptor_dot @LucaPrasso HAHAHAHA METAHUMAN, they are so low quality uncanny valley as fuck

[Luca Prasso @LucaPrasso]

Hint… NanoBanana can modify textures for GNM directly in UV space

[Oliver Prompts @oliviscusAI]

MICROSOFT JUST OPEN-SOURCED SELF-EVOLVING AGENT SKILLS.

it's called skillopt.

skills that improve themselves the same way you train an ai model.

no more guessing whether your prompt tweaks helped. a base model runs the task, an optimizer evaluates the output and rewrites the instructions itself.

→ isolates successful paths from failures to find precise improvements
→ auto-rejects any edit that doesn't beat your benchmark score
→ beats hand-crafted prompts and optimizers like textgrad
→ zero model lock-in, the skill transfers to any model you switch to

100% free. open source.

[Mark Kretschmann @mark_k]

Looks like @AnthropicAI may have done a little distillation for Opus 4.8... from Qwen, no less. Asked Claude who it was, it introduced itself as Qwen.

Pretty embarrassing for a company that has spent so much time throwing stones at others over model training.

People who live in glass houses, Anthropic...

[Corrado Ferrara @CorradoPapers]

@jun_song Ragebaiting, this is so stupid

[Twilly (formerly Twilon) @twilightise]

the ai stories have found league 🥀

[80 LEVEL @80Level]

"Absolutely zero AI was used." SavageCG reworked Blender's Metallic BSDF using 5,403 nodes and data from research papers, bringing over 30 physically accurate metal materials with more to come.

[Vox @Voxyz_ai]

𝘀𝘁𝗼𝗽 telling Claude Code/Codex "no em dashes".
𝘀𝘁𝗼𝗽 telling Claude Code/Codex "stop saying delve".
𝘀𝘁𝗼𝗽 telling Claude Code/Codex "don't sound like AI".

you never gave it a 𝘄𝗿𝗶𝘁𝗶𝗻𝗴 𝘀𝘆𝘀𝘁𝗲𝗺. every README, PR description and landing page ships in the same AI voice, and you ban words one at a time.

orwell wrote the fix 𝟴𝟬 𝘆𝗲𝗮𝗿𝘀 𝗮𝗴𝗼. six rules, 1946. paste them into your global CLAUDE.md / AGENTS.md and every session picks them up:

→ never use a long word where a short one will do
→ if it is possible to cut a word out, cut it out
→ never use the passive where you can use the active

6 blocks you can copy-paste directly 👇