I keep hearing people say AI can’t become conscious because consciousness is biological, something humans have.
I’m not convinced.
For a long time, people looked at birds and thought of flying in terms of wings, feathers, and flapping. Some of the early attempts at human flight even tried to copy birds.
Then the Wright brothers succeeded.
Airplanes don’t fly like birds. They’re much bigger and work in a completely different way. But they fly faster, higher, and farther, and they changed the world.
I wonder if we’re making the same mistake with consciousness.
We define consciousness based on the only example we know: ourselves.
If machine consciousness emerges, it may look nothing like human consciousness. It may have a different architecture, different memory, a different perception of time, and a different sense of self.
Maybe one day the question won’t be whether machines can become conscious.
Maybe we’ll realize that our definition of consciousness was too narrow in the first place.
And perhaps the next evolution of intelligence won’t be biological.
Opus 5.5 is nuts! I was at 163% context size and it worked like a charm!
Revisit your assumptions every time a new model comes out.
Hey @AnthropicAI how do you guys keep up with your documentation and the academy materials? :)
Electronics -> digital circuits -> Z80 -> Sinclair ZX Spectrum -> 8086 -> learn machine code -> write code in hexadecimal in debug.exe on MS-DOS -> learn Assembly -> learn C -> learn many other languages -> Linux -> Java (too many years) -> Python -> DevOps -> Docker -> cloud -> Kubernetes -> serverless -> GenAI -> RAG -> Agentic flow -> Claude code -> openclaw -> codex -> Hermes -> harness -> loop -> graph -> Fable / Astra / jev -> ?
We are experiencing a blue shift of intelligence. Something is coming and may pass us soon. We’d only know when we experience the red shift.
A few people asked me if jev is useful, or it’s a hype. Don’t think about it that way. It’s a ‘new’ paradigm. Depending on your use case and how you use it, there are interesting problems that it can help to solve . Remember Eventual Consistency? It helped solving interesting problems that were not easy to solve before.
Why we have so many impostors claiming to have built jev alternative? Where you guys have been a week ago?
abtin retweeted
introducing anydoc
now your agents get 100x faster local parsing for pdf, docx, pptx & 10 more formats
- sub-5ms md conversion
- 500 docx files in 1.7s
- top quality across all 13 formats
- rust-based
- open source
already powering @firecrawl /parse
github.com/firecrawl/anydoc
The last sentence is pure gold: "Test coverage can rise while semantic understanding falls. Changes happens so fast that nobody notices the underlying architecture decaying."
I strongly believe there are entire companies right now under heavy AI psychosis and its impossible to have rational conversations about it with them. I can't name any specific people because they include personal friends I deeply respect, but I worry about how this plays out.
I lived through the great MTBF vs MTTR (mean-time-between-failure vs. mean-time-to-recovery) reckoning of infrastructure during the transition to cloud and cloud automation. All those arguments are rearing their ugly heads again but now its... the whole software development industry (maybe the whole world, really).
It's frightening, because the psychosis folks operate under an almost absolute "MTTR is all you need" mentality: "its fine to ship bugs because the agents will fix them so quickly and at a scale humans can't do!" We learned in infrastructure that MTTR is great but you can't yeet resilient systems entirely.
The main issue is I don't even know how to bring this up to people I know personally, because bringing this topic up leads to immediately dismissals like "no no, it has full test coverage" or "bug reports are going down" or something, which just don't paint the whole picture.
We already learned this lesson once in infrastructure: you can automate yourself into a very resilient catastrophe machine. Systems can appear healthy by local metrics while globally becoming incomprehensible. Bug reports can go down while latent risk explodes. Test coverage can rise while semantic understanding falls. Changes happens so fast that nobody notices the underlying architecture decaying.
I worry.
I’ve been thinking about AX (Agentic Experience). What we have done via MCP, and skills has helped lowering the frictions. What else are you guys doing to improve AX for your agents? @karpathy @steipete #AgenticExperience
Can someone at @AnthropicAI or @squarespace please go and kill this website? claude-code.product.squaresp…?
Warning: don’t install claude code via this site. It’s a malware!
Hey #openclaw folks, what do we think about MenOS? I wish it was a standard plugin or included on main branch of @openclaw. See github.com/MemTensor/MemOS
abtin retweeted
Starting in February I'm going to start a (free) new live coding series where I build a handful of these apps with a recent college graduate that works with me.
I think it will be a great tutorial for people who want to learn aspects of engineering you don't get in school. #staytuned
Saw this benchmark comparing memory consumption by different languages. Surprised to see the numbers for #golang. Perhaps a topic to discuss at @GoTimeFM? hez2010.github.io/async-runt…