@emollick @bdambrosio is it really far removed from GPT Pro? probably purely on a framing basis.
English
HELPING HOUND // UD.AI // 🐉☁️☁️
2.5K posts
HELPING HOUND // UD.AI // 🐉☁️☁️
@helping_hound
finding homes for unwanted dogs & augmenting human intelligence
https://poe.com/underdogAIguy Katılım Haziran 2009
3.2K Takip Edilen380 Takipçiler
@bdambrosio I dunno, it just looks like Mythos crossed a threshold of usefulness for cyber capabilities. That doesn't require a magical step change, capability gain can just cross thresholds of utility as part of standard progress.
English
I am catching glimpses in my feed that there is a backlash against Mythos as "marketing hype," and it is a little confusing.
I don't think anyone who has used the latest agentic coding tools, would think that expecting large-scale cybersecurity implications of increasingly good AI models is unbelievable, especially after reading the red team reports. It feels like a better place to start is to assume that there are new risks, and then we can all laugh at Anthropic and pat each other on the back if there are not.
Also, while the AI labs certainly are impressed by their own accomplishments and benchmarks are flawed, I would note that both publicly and privately, Mythos seems to be taken seriously at a lot of large institutions and organizations filled with smart people who would rather not be worried about a new cybersecurity risk.
Finally, I am not sure "our product is dangerous and we need to alert the government to that" is the sales pitch to the corporate world that critics seem to think it is.
English
@hammer_mt @dbreunig it is kind of inevitable that each use you push it further, adding slightly more complexity, context length or unfamiliarity for the model causing it to hiccup - I do think that's what most people experience even if not all (they're obviously always trying cost cutting measures)
English
@dbreunig I'm not a conspiracy theorist and I like to think I'm a fairly sophisticated user, but I talk to Claude more than most colleagues and it's immediately obvious when something changes. It's like a friend coming into work drunk or with a concussion.
English
Every time I see people asking if others are seeing a Claude/Codex/etc regression, I wonder what they’re building and prompting.
English
@JeremyNguyenPhD arxiv.org/html/2601.1805… think this explains it
English
Scrolling is pure evil. An hour of brainrot doesn’t just leave a hole where something meaningful could have been but also actively degrades the machinery you’d need to fill that hole.
It corrupts your capacity for sustained attention. Books become harder, conversations feel slower, your own thoughts start to bore you. Over time, the range of things that can hold your interest narrows until you’re left with a shrinking circle of stimulation that only the algorithm can satisfy.
It erodes your relationship to yourself. Curiosity fades. Compassion requires a kind of patient attention that atrophies. You stop wondering what you care about because the question itself feels effortful. What’s left is a stable, “comfortable” numbness. Not only five years subtracted from your life, but a slow hollowing out of the person who would have lived them.
Justin Skycak@justinskycak
You have to understand. Spending 1 hour per day on brainrot is insane. That's about 6% of your waking day. About 5 years of your waking life. Half a decade. On brainrot. Just gone. Zero return. Zero fulfillment. Zero meaning. Zero contribution to the other parts of your life.
English
@Only4CBN @zephyr_z9 south Korea ETF has a large amount of exposure
English
@mf_tomb @MurrayHillGuy1 the optimal time saving strategy for men is also often to match more and talk less
English
@MurrayHillGuy1 It’s because too many women just use the app to farm attention and rarely meet guys off there
English
Hinge just launched a new “Signals Badge” that basically lets others know whether you ghost or not…wtf is this shit?
English
I will pay 20% management fees
for allocation to
A n t h r o p i c
S e c o n d a r y
S h a r e s
~~~
~~~~~
~~~~~~~
~~~~~~~~~
~~~~~~~~~~~
~~~~~~~~~~~~~
~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
English
@Vexian you have to be particularly chopped to actually watch live streams
English
Short form clipping and leveraging distribution channels.
Just a theory but it a bit of a house if cards at the moment.
wally b kleeposting@kleeposting
What I don't understand about zoomer celebs is that they're all over the media but when you look at their actual streams it's less than Northernlion
English
and when ai does this we call it a hallucination that makes it unusable
Kpaxs@Kpaxs
The world literally reorganizes itself around what you've trained yourself to notice. A doctor learns to read symptoms. An architect learns to read space. And once you learn to see something, you can't unsee it. The architect is forever cursed/blessed to notice bad spacing. The editor sees typos on billboards. The musician hears the out-of-tune note that nobody else catches. Education is about installing new filters, new sensors, new ways of parsing reality so the noise become signal. And the most interesting people are the ones who've stacked multiple lenses. They see connections that specialists miss because they're triangulating from different vantage points.
English
Czech automaker Škoda created first bicycle bell to penetrate noice-cancelling headphones.
They tested 100s of headphones and found a low frequency (750 to 780 Hz) that anti-noise algos can’t suppress.
Bell is mechanical and uses an irregular hammer mechanism.
This might be the most European tech story ever and feels inevitable that an audio firm will create an anti-anti-noise-cancelling-bell headphones.
***
Link: youtu.be/zDaVPfpQvPI?si…
YouTube
English
HELPING HOUND // UD.AI // 🐉☁️☁️ retweetledi
Allowing Sonnet to "phone a friend" (i.e. call Opus) increases performance while also reducing total cost since it reduces tokens spent trying to solve more complex tasks
Claude@claudeai
We're bringing the advisor strategy to the Claude Platform. Pair Opus as an advisor with Sonnet or Haiku as an executor, and get near Opus-level intelligence in your agents at a fraction of the cost.
English
HELPING HOUND // UD.AI // 🐉☁️☁️ retweetledi
Good communication skills required to truly Ai-maxxx
Thariq@trq212
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
@fabianstelzer @xlr8harder au contraire the recent follower tab has finally got me to stop using "for you"
English
@xlr8harder following mostly doesn't do much anymore, same with reposts / quote posts, so not surprising to see people use it less, esp lurkers which is 90%.
it's still useful if you want to engage with someone directly as mutuals through DMs, which is the most underrated feature of X
English
@Observer_ofyou @nikitabier try Pokémon impressions instead
Français
Can't wait to hear what everyone sounds like
XChat@chat
Voice Notes on XChat are finally here.
English
@GoldMagikarp42 @wagslane slower to get them to edit but fast to get them to write the first draft and then edit yourself
English
@wagslane Also it’s slower to use the models to edit than do it by hand. When you ask them to adjust framing of a topic, they never remove the old frame, they rewrite the new frame as a reaction to the old frame. Guess what happens if you try to fix this framing?
English
Anyone that thinks AI is even REMOTELY good at writing is a terrible writer
I've been giving SOTA models pages and pages and pages of my writing for the last 3 years as examples, and they still can't produce a single paragraph that I'd be comfortable putting my name on
English