Cofounder https://teapawt.org/t.co/hOrykl2OuQ: Your marketing command center. The AI that learns what works, then does more of it.
London, UK
Joined March 2009
- Tweets12.4K
- Following386
- Followers514
- Likes11.2K
Pinned Tweet
USA vs UK
Green dots = website visitors with high buying intent.
Yellow = maybes
Grey = not interested
USA is 6 times the population of UK. Market size matters.
when iphone etc hardward allows instant smooth translation, including interpreting cultural norms, the amount of business some countries will suddenly get...
right now it's just not smooth on phones. It's good but the whole press button, then other person presses button thing... it works but it's not human smooth.
troublemaker in shenzen going to rent larger space because of how many people are turning up to build hardware in shenzen.
already teams from san francisco there.
investors arriving too.
not quite vibe hardware yet but given how many companies in that region can build what you want, it's closer to it than any other part of the world.
strain wave drive manufacturer Laifual.... selling mostly to humanoid robotics makers? seems odd
Thought it would be manufacturing but there must be a lot of humanoid robots being made over there, unless there was a miscommunication with their rep.
Shenzhen Robot Expo shows a couple of things...
First of all, there are no humanoid robots walking the streets of this city, which has some of the most high tech companies. There are drones... but even those are few in number.
The robots at the convention were largely remote controlled and can do the kicking and jumping and everything else, but for precision tasks you had to go to the manufacturing hall, and everything there was just older upgraded manufacturing tech with AI added (or not)
However, within that manufacturing tech, you can imagine there are going to be immense improvements.
For every robot, there were ten times as many other products, more focused on manufacturing.
In the streets you don't need Robots because there are so many people willing to do the jobs that you've got graduates as baristas in many places.... and robots could not do those tasks as well or as fast.
It looks like it will be some time before you've got a proper household robot doing that range of tasks.
I see how some people can watch flashy robot demos on the timeline and think with a shiver that the robot apocalypse is around the corner, but reality check:
There is no robot in the world that can change the batteries in a jack-o-lantern.
If you have a robot friendly door handle it might be able to walk outside, pick it up, and carry it back inside. The robot could open the drawer with your batteries in it, but then things get tough.
Opening the package of AA batteries?
Extracting a precision screwdriver from its case?
Unscrewing four tiny screws that require a hard press and twist?
Popping batteries out of the case?
Yes, you could build and train a specialist robot to do that today, but there is a very long way to go to replace the full range of human capabilities. Near term doomer scenarios should stop at “secretly controlling humanity” instead of “wiping humanity out” 😀
can do immense damage to open source with even one model pulling some weird stuff like this.
don't know if true or not as there's an open source vs everyone battle...
don't like any of this
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.
Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.
Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.
But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?
We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.
Read the full report: cohere.com/blog/concerning-b…
Are china working hours for tech companies 9 am to 9 pm at six days a week?
Spoke to our guide about tencent offices and the other big tech offices here in Shenzen and he said, at around 9pm, there are swarms of Taxi drivers taking people home.
Chef in nearby restaurant also said the people coming to lunch barely can communicate 😂. He was describing the hard core nerds.
Shenzen china.... don't see any humanoid robots walking around.
But then, there are so so many taxis and scooter delivery folks (and a few drones but very few relatively)...that the humanoid robots are not needed.
Most of that AI robotics is probably in manufacturing.
Interesting note from chinese guide is that the barristas and other folks are often graduates. Labour market ain't great for youth.
What is great is the scooters and vehicles are electric. It's silent and free of fumes.
Scooters are near universally running on the pavements.
Clean and safe...
17 million people.
Someone we spoke to mentioned pre-covid china vs post covid has resulted in change. Pre-covid it was friendlier and post covid it feels worse... not sure what to make of that.
All in all, safe and clean is better than dirty and dangerous.
Easy to forget how much of an upgrade this is for so many people in China over the decades.
these japanese motherf**kers. They made a better egg mayonnaise / salad sandwich than we've ever had.
They made our sandwiches better than us.
Kullar retweeted
We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks:
Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better:
Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better:
Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better:
Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work!
In summary:
- As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding.
- Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.
Kullar retweeted
The go fund me page of guy who leaked the City emails.
They're stripping his security.
Donate or don't donate but share it at least.
On Wednesday, just hours after the verdict officially confirming Manchester City’s guilt, the Commission responsible for Portugal's witness protection programme informed me of its decision to terminate my participation in the programme and withdraw my police protection, thereby knowingly placing me in a life-threatening situation.
Given these circumstances, I have no alternative but to relocate, within the next few days, to an undisclosed location for an indefinite period of time, due to the multiple death threats I have received and the serious risk of an assassination attempt against me.
For this reason, I have just set up a GoFundMe campaign and I am asking for everyone's support to help me survive this serious situation created by the Portuguese State, which, as I have already stated, has knowingly decided to place my life at risk.
I would like to take this opportunity to extend a special thanks to everyone who continues to support Football Leaks, and I can give everyone one guarantee: the fight will most certainly continue!
gofundme.com/f/help-save-rui…
I'm sorry but what?
Company founded in Sept 2024, sold for 8bn 2 years later.
Smart AI folks are making bank, though to be fair, Fei-Fei is part of the reason deep learning kicked off...
Apple’s version of grokbot / muse going to win? Integrated into everything already including via airpods and their own version glasses.
Big companies just get bigger!
