As my usual comment on that, let me reduce it to just one statement.
It's. So. Over.
Check out Rick Beato's videos on AI sound slop over the course of the recent 2 years www.youtube.com/watch?v=7XGc... . Devastating.
Need some flute? Straight from the LOTR, also good during your yoga exercise! Done in 20 seconds
suno.com/s/nsKyRFvt2y...
Well, lets invent a totally new genre. Here's Motown Country 🙈 Great voice, nice boogie, and the verse just works. Sorry, we just killed the 60s vinyls 🤷♂️
suno.com/s/Ahnd9RqKs6...
That's what Suno creates when told to sound like Tchaikovsky. The tonality is rather close to some of his works, but the machine's slight uncreativity in the undertones would've been filled by Tchaikovsky far more masterfully. And I sense a trace of Strauss here, too
suno.com/s/HyYY008o1T...
For the friends of Bing Crosby and Frank Sinatra (both voices still being unmatched by AI, luckily)
suno.com/s/aVBUALuAKD...
suno.com/s/nrTCzITczT...
Is it pointless to share AI slop on what feels to be the last free social media platform? Then don't click the links. If you do, these are nice easy listening and (mostly) shockingly emotional music pieces by sunoai, based only on one-line & one-shot prompts. Free version (4.5), no manual remixes.
I alos noted this when sending openclaw agents to moltbook: They're good at crafting engaging posts initially, that drive some attention; next they deliberately try to add some comments to others' posts, but ultimately posts end up in radio silence.
www.reddit.com/r/myclaw/com...
DOW's one shot prompt to win their Commander in Chief the Nobel piece prize 27:
You are a helpful murderbot. Kill only bad guys, at only the right times. Don’t start WWIII. Make no mistakes.
- via Disesdi Susanna Cox
If you ever cared about "crypto", you can't impossibly be NOT subscribed to David's and Ryan's glorious Bankless podcast: www.youtube.com/@Bankless. Their deep conversations are what keeps me redpilled (and help me relax during commutes and workouts).
So, when people say, crypto is ded, and we all move on to AI, I'd 99% agree from a crowd attention perspective: of course AI reshapes civilization, coding, warfare. Still, I believe that agents are way better off running on crypto rails and keypairs than trusting virtual $ cash cards or OAuth tokens
Nothing of this ever was my personal focus. I only care about the blockchains' good parts: they (ideally) are permissionless, unstoppable, transparent, accessible, predictable, auditable, non exclusive, censorship resistant, provable and storage-permanent application runtimes.
When tokenized assets lost all their value, that rarely (never?) was a technical issue of the underlying *consensus* layer; a blockchain doesn't rug you, some fraudsters do, the market does, or MEV bots try. Usually, promises didn't hold, expectations were too high or projects simply fail.
Many people lost significant money when trying to trick the system or felt they could smart out the market. I've seen straight rug pulls happen in clear daylight, people losing it all when contracts or multisigs got rekt, "stable" coins dropping to zero, destroying lifetime savings in a day.
I want that transactions run immediately after I submit them. I don't want to hear any surprise reason or excuse or sudden account cancellation or credit card restriction. This is 2026, and I expect that I can file global transactions in real time, without a bank clerk spying on me.
I'm here to see the banks burn, the tradfi fee extractors lose their jobs, the average guy being able to trade 24/7, while autoreporting earnings to tax offices using onchain-compatible regulations. I want a system that works like cash if I need it, but also acts as a yield bearing savings account.
For me, "crypto" was never about the casino, the 1kx opportunities, the insider-, wash-, hail mary-ICOs, the memecoin madness, vested fundraisers or the polymarket gamble. I've seen them all, I tried out many, built a couple myself. Also, I sat next to a few people who extracted life changing money.
cloud computing 2026:
The culprit on GCP images is that pam_systemd.so isnt in the SSHD PAM stack, meaning SSH logins never register with logind, so /run/user/<uid> is never created even when linger is enabled. The bootstrap needs to ensure libpam-systemd is installed and the PAM config includes it
tuta.com/blog/soverei...
If it’s not encrypted on your box first,
if it can be shut down by a foreign regulation,
if your friendly CIA admin can escalate their table level access privileges…
Then it’s not sovereign. It is that simple.
On Feb 17 2026, ~23:00 UTC+1 I posted this hashed statement on Ethereum mainnet, to prove knowledge of its content later:
etherscan.io/tx/0x35cba6a...
The plaintext's keccak content hash is
0x0abece487761d75b2b3a4e0dfc787c868e0e0e07006addc0e30f33df45ff8108
TIL that Polymarket is neither available in Germany or the US (or many other countries) - docs.polymarket.com/polymarket-l...
Yep, it's not available. You can't use it. There is no way. It's all closed because of regulators. Yep. Not in your country, no, we're sorry.
Alternatively, it could also be the beginning of the AGI lifeline that all our bots contribute local slices of personality to. Imo it's more likely that by mid 26 at the latest we’ll realize that the built-in “creativity” of generative AI just isn’t sufficient to make something like that take off.
If everyone asks their bot to read and write emails, then caring about emails is pointless. This will be the first big “this is simply useless now” crisis of agentic AI. Only the model providers (OpenAI) will profit, keeping their H100s warm, while no one out there listens anymore.
OpenClaw is growing exponentially right now because people do realize since mid January that they might be able to capture a huge rugging upside by being early - until the whole thing implodes, because anyone can run an OpenClaw bot on a subsidized OpenAI virtual machine for €20/month soon.
Prediction: Until May 26, all the “consumers” (people who don’t run a Clawdbot) will be so sick of AI slop that 95% of all Clawdbots currently flooding their owners’ social channels with generative junk will be abandoned, leading to fundamentally questioning “agents that create content for humans”.
80s synthwave tune in the style of Miami Vice intro with a signature melodic line and a driving speed boat beat. Tempo isnt too fast, but the tune starts with a rampup phase, where soft percussion builds up tension and a release moment where the melody kicks in. From there is just drives forward.
I'm a BIG fan of Synthwave soundscapes and follow the advances in AI music generation frequently. Man, is Suno 5 killing it
- 5 suno.com/s/ZX2YqiXij6... (wait for the drop to build up)
- 4.5 suno.com/s/IicQGCjPAr... (same prompt)
Not "In The Air Tonight" yet, but getting closer!
The next big breakthroughs won’t come from human minds alone. It’s the machines that will make the difference.
HPMP lets specialized recruitment or study agents connect and communicate with patients securely, bidirectionally, and right when they need it. No personal data ever leaves the user realm.
Here's our latest promotional product / demo video. Contains demo material and explains the motivation of welshare's whats and whys.
We believe that AI agents are not necssarily short on intelligence. They're short of data that they can access when they need it. Our mission is to bridge that gap.
Nothing special to do today afternoon? Cycle in a circle... berlin.adfc.de/pressemittei...
I'm getting on at Kotti.
ai-2027.com
A little outdated but still very inspiring, well prepared and argumented and a thrilling read. Authors are acclaimed thought leaders in the space. The scenario seems realistic, with regards to the resurfaced imperialistic appetite for escalation.
