This happens when you don’t patch you apps: some tiny dirty hack happened on my server, installing a krypotminer named kinsing, by using a vulnerability in log4j. I noticed some strange CPU load and an execubale named kinsing in /etc . It instaled itself in crontab through a system entry when re-starting. And uhm, yes, tha application using log4j had root privilieges… I know….
Aber McKinsey ist nun auch einmal kein ganz kleiner Laden und somit können die sich schon aufwändige Studien leisten, die oftmals einen wahren Kern haben.
Nun also die Veröffentlichung der Studie zum ökonomischen Potenzial von generativer AI.
Die Key-Takeaways sind:
Von den 63 Usecases, die McKinsey analysiert hat, geht ein Produktivitäts-Gewinn weltweit von ca. 2.6 – 4.4 Billionen Dollar erwartet. Das Bruttoprodukt der UK war im Vergleich in 2021 $3.1 billionen. Dies sei aber nur ein Anfang. Über welchen Zeitraum hier gesprochen wird, bleibt unklar. Die 63 Usecases finden sich unter o.A. Link wieder.
75% dieses Wertes entstehen in den Bereichen Customer Operations, Marketing & Sales, Software Engineering und Research&Development.
Alle Branchen sind betroffen, vor allem aber wohl Banking, Hightech (was auch immer das sein soll) Life science. Aber auch der Handel soll mit ca $400 – 660 Mrd. profitieren.
Einer der größten Hebel liegt aber in der Veränderung der Jobs: 60-70% der Arbeitszeit, die Mitarbeitende auf wiederkehrende Tätigkeiten verschwenden, werden durch die Möglichkeit mit natürlicher Sprache (Natural Langauge) zu interagieren und diese auch auszugeben, eingespart! Das ist eine Menge. Dies betrifft insbesondere higher paid jobs, sog. Knowledge Worker.
Bis 2045 würde die Hälfte der jetzigen Tätigkeiten automatisiert. Ca. 10 Jahre früher als bei bisherigen Schätzungen.
Wir stehen mit der Veränderung am Anfang – Unternehmen und Politik haben noch Zeit zu reagieren (klar, sie sollen ja auch McKinsey beauftragen 😉 )
Gleichzeitig werden in der Studie natürlich auch die Herausforderungen genannt, die jede generative AI mit sich bringt – diese werden allerdings nur stichpunktartig erwähnt:
A new study was conducted regarding the results of different language models. The main outcome is: size doesn’t always matter.
Large lang models (LLM) are trained on up to 530 billion parameters which results in significant cost effetcs. The study shows, that models with much smaller parameters like chinchilla (70 billion parameters) outperform ther colleagues, especially when raising training tokens.
This is the 5 Shot performance of differrent models
The conclusion we can draw from this are:
it is indeed possible to use only publicliy avalable data to train a perfectly working language model. AI is going to stay, regardless the licensing-wars we will see with OpenAi etc.
It is possible for companies to add their own „language“ to existing models at a doable pricetag
You should not stick to one model, buzt be flexible and interchangable with the results by testing, testing, testing.
(Dall-E prompt for header picture: "Create a picture where the language model "Goliath" is being beaten by the language model "Chinchilla", make that a fantasy picture and Goliath being a big, fat bear, as where Chinchilla is a very strong mouse.")
Nico hat vor ein paar Tagen drüben bei LinkedIn die steile These aufgemacht, dass AI nicht als Abkürzung für Artificial Intelligence, sondern für Augmented Intelligence stehen sollte. Und damit hat er aus meiner Sicht komplett recht und einen wichtigen Punkt: Künstliche Intelligenz sollte „nicht in einem bedrohlichen Kontext gesehen werden“, sondern eher als „Erweiterung unserer geistigen Möglichkeiten“. Das geht dann in die gleiche Richtung wie vor 30 Jahren: Wenn es Wikiepdia gibt, warum sollte man noch irgendwas auswendig lernen – man muss nur wissen wo es steht.
Victor hat dann das Thema Infrastrukur in die Diskussion eingebracht, die ich sehr spannend finde. Es sind ja im Moment die Modelle wie OpenAI und Bard, die mit ihrer Leistungsfähigkeit für Aufmerksamkeit sorgen – aber die Implementierung in den Business Kontext (wenn er das meinte) sehe ich noch nicht. Ja, Microsoft bringt natürlich mit Azure eine 0365 nahe Entwicklerplattform mit sich – und Copilot integriert OpenAI in verschiedene Officetools. Aber ist das dann schon das gelobte Land? Viele Unternehmen werden sich fragen: 1. Was passiert mit dem ganzen Silowissen in den Datenbanken, FAT Applikationen. Wie kommt es da raus und wird für meine Mitarbeitenden und Kunden nutzbar gemacht?
2. Wie sieht das Interface aus? Ist es wirklich der Chat, der als virtueller Assistent „neben“ mir steht?
3. Wie messe ich die Performance der AI im Unternehmen?
4. Wie stelle ich den Wahrheitsgehalt fest und mache ihn transparent?
5. Gibt es genau EIN LLM das zu mir passt?
6. Und wie integriere ich die Business-Prozesse mit der AI?
Denke, da ist noch Platz für viel Infrastruktur um Augmented Intelligence für Unternehmen wirklich sinnvoll nutzbar zu machen.
I’ve been an avid runner since years, doing some races, marathons and triathlons. I would say, that in general I am aging well, but since I became 42 or so, my knees started to hurt when moved in awkward ways. It’s not always but from time to time it hurts so much, I can’t do another step.
Some month ago I came across the book „Born to run.“ It’s about some south-american tribe, which up until today are running barefoot like 120 miles a day 😳. It’s not a classic runners book, more like a novel, and it gets to the point, that running barefoot is the „natural“ way of running (like in ancient days, chasing an anthilope to death), using the architecture of our body as a natural „spring“. They say it prevents injuries and typical achers in the achilles heel and so.
So I got myself a pair of these:
Vibram fivefingers kso evo
Yes, they look stupid. They look stupid when I wear them. But they feel sooooo good while running. Not in the beginning though – but step by step:
My first run was about 15 minutes and my calves were exploding.
My second run was 25 minutes – and my calves were exploding.
My third run was 75 minutes – and I felt nothing than pure pleasure!
Running „feels“ so more light and natural; with every step you sense the underground.
I’ve now been running the fivefingers since 3 month and don’t want to go back. Seriously, it changed my whole running stile. I am more aware of and in the run, my heartrate is way lower. I am bit slower though, and especially on bigger gravel I have to watch out where I set my foot (it helps to set smaller steps).
From time to time I am in my nikes, but they feels clumsy and awkward now. (btw.: Nike invented the „cussion“ running shoe, to sell more pronation-correcting sportshoes, leading to more injuries – „true“ story 😉
Pro tip: as you run barefoot in your barefoots: vinegar and shampoo in the sink give wonderful results regarding the „smell“ effect 🙂
Today 5 years ago I was driving on the german Autobahn A5 at 05.30 in the morning, when a sleepy van driver changed onto my lane unexpectedly I had to dodge it and crashed into a lorry nearly unbreaked.
Thankfully I was not injured but my car was totally destroyed as I had hit the lorry with about 130km/h.
A moment that changed my life. This is what I took from it:
Big german cars save lifes – literally 😉
Always expect the unexpected
At a certain speed, physics is just physics – no software or hightech can help. When life hits, every mm crumple zone counts
After the crash I was not driving for 4 weeks. I was too afraid. Eventually I got back into the cockpit and there is no anxciety anymore. But my average speed is down significantly.
The other day I was drawn into the retro wave and got out my old iPod classic. I still love the iconic interface which set new boundaries.
I especially like that is has no internet connection and thus no notifications or other connectivity which distracts from – well, just listening to music.
On the other hand its annoying, that you have all these subscriptions: Spotify, Apple Music, Deezer, and can’t download the songs to your iPod directly. Enter python and spotdl.
Once you download it, and install ffmpeg addiotionally, you feed it a link to either a spotify song, album or playlist and it translates this to a youtube music link. ffmpeg then grabs the link and meta data like album art and converts it to 128kBit mp3 on your HD.
Is this legal? I am not sure. And I don’t care. As I do have all these subscriptions and don’t share the downloaded mp3, I see it as just another channel on accessing what I am paying for.
I was using LittleSnitch years ago and installed it again yesterday. LittleSnitch is like a neat little firewall (basically it’s a socket filter for macos) – but it also now has this neat map in a great design showing at a glance, where your data is going to geographically:
Where is my data going to?
Its funny to see, how some apps are even sending data home where they were not supposed to. I don’t get why deepl, my favourite AI translator, hosts its data in the US, wheras its just a kilometer away from my office in cologne. They should not.
LittleSnitch comes in a „silent mode“ which by default allows all traffic, but sends you a notification as soon as a „new“ connection is established – and you can then decide if you want to allow or forbid that.
As I mentioned previously, ChatGPT is quite good (but not only at) at SQL. I think its a great opprotunity to really learn to code (if you want to call SQL querying „coding“ but that’s another discussion). I mean, only this information is SO valuable, I would have searched stackoverflow for hours finding the reason for my SQL bug:
I now can change the query or update mysql.
So here is my lazy setup to create complex SQL queries:
I am using the graphical query builder of metabase for all the complex joining of data
2. I then review the result and convert the results to SQL.
3. For the lazy mode I then paste the sql to ChatGPT, and asking it to add modifications, adjust it – and all in natural language.
Having binged The White Lotus season 1 I have to say, it reminds me of a long-format Instagram movie. Filters being applied to the scenery are just beautiful, but I find them sometimes a bit too much. The soundtrack, though, is incredible. It got 10 emmys recently.
Update: I found series 2 even better than series 1 – it’s more complex and the ending is…. 😉