Had not seen this one:
"AI’s Ostensible Emergent Abilities Are a Mirage"
with the whole "#AIs will magically become more competent with more data because emergence" thing coming up now that the capabilities of AI systems are mediocre for most real world cases this is good to have as a link.
It's my pleasure to share with you that #Trajectools 2.0 just landed in the official #QGIS Plugin Repository. This is the first version without the "experimental" flag. If you look at the plugin release history, you will see that the previous release was from 2020. That's quite a while ago, and a lot has happened since…
@underdarkGIS Here are some ship trajectories entering the Bosphorus from the Black Sea… colour = port of origin. Thanks for this amazing tool! #QGIS#shipping#AIS
I’m continuously testing the algorithms integrated so far to see if they work as GIS users would expect and can to ensure that they can be integrated in Processing model seamlessly.
The distribution of values means a linear gradient only shows the busiest of sea lanes. So I used log10(log10(value)) to normalise values with #qgis raster calculator.
I know there are plenty of folks here who are #adsb nerds but what about #ais? Are there open AIS networks like that have been built for capturing ADSB data?
...on #embodiment, and several companies are working on that front. The very basic reason being 3-fold:
IMO real-time and sensory data access is necessary for consciousness
several (non-computer) scientists agree that a corporal, physical presence is a step in evolution.
The moment an #AI is hooked up to sensory data world wide, it will be a tremendous amount of data very different from texts and videos. Once it learns to use it somewhere, all connected #AIs know this at once.
There are lots of general ideas about how to utilize #LLM#chatbots in practice, beyond just chat.
Using tools, plug-ins and APIs.
Ad-hoc programming, running the programs to perform actions which would be otherwise hard for chatbots.
Vector embeddings, indexing knowledge more effectively than ever, retrieval augmented generation.
Multi-modality.
Autonomous execution as agents. Memory. Societies of agents. Graph of thoughts, algorithm of thoughts.
Combining ontologies and knowledge graphs to chatbots.
Fine-tuning, quantization, distillation into smaller models to run them cheaper and in small edge devices.
What else? Where to go from here?
I would say we will soon get to "zero-shot on steroids" as we are training LLM chatbots in self-competition now in many research groups. They will still be data starved. There is a lot they don't have in their training corpus which is valuable for them for creating value.
How to get this knowledge to them? We need a way where the agents will actively seek people who know, chat with them to learn what they know over a long, open-ended time period like a student, and encode this knowledge somewhere in an easily usable form, for example in a RAG knowledge base.
We need organizations to transform into schools for #AIs. Where they are put to work but where there are life-long experts, gurus, available as living books for them to get that hard to reach silent knowledge out and shared out. We need to reinvent the human role, not as laborers but teachers.
This type of "work" probably never ends, even though its marginal value would seemingly go down as at some future point the most important nuggets of gold have already been taught to the machines, but the infinite curiosity and near limitless capacity to multiply value out of any tiny tidbit of knowledge will make it valuable to have people around in AI organizations, even if not for work, management and leadership, but at least as advisors and as someone to chat with.
We should already start building silent knowledge exraction systems driven by chatbots as backbones of our organizations to be ready for the future. This knowledge not yet written in books nor crawlable in the web will attract AIs and AI capability and power around it — it is a fountain of knowledge, and as such forms an oasis where intelligence grows.
People are worried about who to hire now when even complex coding tasks are easily solved by #AI. This worry is natural and good, and based on great instincts.
In the past you hired people you expect to stick with for 5-10 years, to scale up your organization's throughput and capability.
It's different now. The jobs won't exist in 2 years anymore. Should you stop #hiring? NO!
You must hire to get there! You won't get there is you passively wait. It's a different place, you need rowing to get there.
You will need to transform your organization into an #AIOrganization and you will need to hire people who have skills they can teach to #AIs. You don't hire people to do the job as much as you hire them to teach you AIs to do the job.
Det växer verkligen på sommaren. I ett hörn av tomten t ex en hyfsad AIS-antenn o ett litet skott ADS-B-dito. Vid roten en låda m en raspberry pi powered over ethernet som kör PiAware. Matar flightaware, flightradar24 samt marinetraffic #raspberrypi#ais#adsb#poweroverethernet
Today's #AIs are really just large statistical models gobbling up billions of data points from all over the internet and regurgitating them without understanding a thing. Like parrots. Super parrots. Very expensive ones.
IMHO it works well in conversation. When somebody says "#AI will take your jobs" you respond "Oh yes, all those billion-dollar super-parrots will do a much better job than me."
@coupland False analogy. I'm only complaining about the current crop of super-parrots. I am quite certain some #AIs can be built that aren't super-parrots, except that that's not the current crop.
#Russia is faking marine AIS data on Black Sea, painting a giant Z letter near Crimea coast, reports UK MoD. The ships that painted it with their transponders were reporting moving in zigzag pattern at 200 km/h.
Almost every week now, + despite statements to the contrary, by many #AI#scientists and #programmers, the utopias of #IsaacAsimov and #PhilipKDick (+ others 1)) are making a leap forward.
Due to all the white noise + the hype regarding #AI most of the general public.
Even without #AGI, "...the current crop of less sophisticated #AIs already poses a threat, from #misinformation-spreading synthetic content to the threat of AI-powered #weaponry...
Although #humanity is racing toward a cliff, we're not there yet, and there's still time for us to slow down, change course and avoid falling off – and instead enjoying the amazing benefits that safe, aligned #AI has to offer,"..."
(4/n)
"...UPDATE:
The #AIS data for the #Swedish warship that appears to have rushed to investigate something at the site of the NS explosions on September 22nd is no longer visible on #MarineTraffic. If you search for the #MMSI number on that date now, it now shows no data available. I won't rule out a strange bug or other weird cause, but the fact that it was still available on the 25th and then gone on the 26th, the day after this piece was published, is very odd."