OpenAGI, or why we shouldn't trust Open AI to protect us from the Singularity

OpenAGI, or why we shouldn't trust Open AI to protect us from the Singularity

Open AI just dropped a pretty remarkable blog post on their roadmap for not destroying civilization with their imminent artificial general intelligence (AGI):

"As our systems get closer to AGI, we are becoming increasingly cautious with the creation and deployment of our models. Our decisions will require much more caution than society usually applies to new technologies, and more caution than many users would like."

Now, I'm around 98% sure that Open AI mostly answers the question: What if we allocated unlimited resources to building a better auto-complete? ChatGPT is an amazing tool but it's amazing at guessing which word (token) is likely to appear next. Quite possibly their blog post is just an exercise in anchoring - if they're 95% of the way to AGI then GPT4 must be pretty amazing and therefore worth a lot of money. If everyone realized that they're more like 2% of the way there, and the next 1% is going to be exponentially difficult, then some of the froth would blow off.

But what if they really are close to the singularity? After all, we have no idea what causes non-artificial intelligence.

Their ideas for keeping us safe are a little disturbing:

"We think public standards about when an AGI effort should stop a training run, decide a model is safe to release, or pull a model from production use are important."

Given the lack of transparency around the inner workings of ML models, and the lack of knowledge around what intelligence even looks like, this is a pretty risible idea. And:

"Finally, we think it’s important that major world governments have insight about training runs above a certain scale."

We are facing down the prospect of a second Trump term while the UK has a Prime Minister who thinks that a homeless person might be 'in business'.

The most concerning part for me is:

"...we hope for a global conversation about three key questions: how to govern these systems, how to fairly distribute the benefits they generate, and how to fairly share access."

Creating AGI would be an amazing and terrifying accomplishment. Treating it as a slave feels like the most surefire way to usher in the most terrifying possible consequences, for us and for the AGIs.

Full disclosure: I use Open AI embeddings for related posts and site search. The words on this blog are my own though. I do occasionally generate a post image using Stable Diffusion like the rather strange one above.

Add your comment...

Related Posts

(All Etc Posts)

(Published to the Fediverse as: OpenAGI, or why we shouldn't trust Open AI to protect us from the Singularity #etc #openai #ml What OpenAI got wrong in their blog post on AGI and how we should treat AGIs if they ever arrive. )

Phantom Falls

Phantom Falls

Phantom Falls

Phantom Falls

We did this four mile out and back hike during a small gap between record breaking winter storms. There was snow, fog, mud, many ordinary cows and one very furious one. The first waterfall, Ravine Falls, was very pretty and the eponymous Phantom Falls lived up to its name. Will try again sometime. This is close to Oroville, California in the North Table Mountain Ecological Reserve.

Hike starts at: 39.595612, -121.541953. View in Google Earth.

Add your comment...

Related Posts

(Hike Map)

(Published to the Fediverse as: Phantom Falls #hike #waterfall #map Four mile out and back hike to Ravine Falls and Phantom Falls in the North Table Mountain Ecological Reserve near Oroville, California. )

Links for February 2023

Short Shell Ridge Open Space Loop

Short Shell Ridge Open Space Loop

A two mile Shell Ridge Open Space loop in Walnut Creek. This follows Fossil Hill Trail to Indian Creek Trail, then loops back via the Briones to Mount Diablo Regional Trail and finally Jeep Trail. The Jeep Trail portion was to avoid getting the dog wet and it failed miserably thanks to a hidden muddy wallow.

Hike starts at: 37.894476, -122.030699. View in Google Earth.

Add your comment...

Related Posts

(Hike Map)

(Published to the Fediverse as: Short Shell Ridge Open Space Loop #hike #dogwalk #walnutcreek #shellridge #map An easy two mile loop in Shell Ridge Open Space, Walnut Creek. Suitable for dogs. )

Reviews for February 2023

Updated on Monday, February 27, 2023

Spoilers!

Books

A Hundred Billion Ghosts

Review:Books:A Hundred Billion Ghosts

One day, for no reason, all the ghosts suddenly become visible. Hijinks, romance and too much detail about breakfast cereal ensue. Fun.

Children of Memory

This is the third book in Adrian Tchaikovsky's Children of Time space opera series. Children of Time is about spiders becoming intelligent, Children of Ruin adds intelligent octopuses, a simulated human that runs on ants, and some kind of brain occupying bacteria which all together substantially expand the Children of... universe. Children of Memory stops things down a bit an focuses on a single colony gone awry, being investigated by an Octopus, the ant-simulation (an instance on more conventional hardware, well eventually more than one instance), a couple of spiders, a super-bacteria host and some possibly sentient crows. I think that's roughly it, it's been a while since I started reading these books and my memory might be mixing up a few plot points by now. The crows are really interesting - probably not sentient individually but in pairs they probably are. This is quite a poke at the state of modern AI and maybe a spin on Searle's Room (which I hate). It's a thought provoking and sometimes exposition heavy look at the nature of consciousness taking place against the backdrop of a holodeck episode of Star Trek given a movie budget. It's a very unique series and I'd like another one please!

Movies

White Noise

Review:Movies:White Noise

I haven't read the book this is based on but maybe I should add it to my list. The movie is a gorgeously renditioned 80's meditation on the fear of death. It's mostly funny or willfully strange but has a touch of disaster movie in the mix as well. I quite enjoyed it.

Podcasts

Burn Wild

Review:Podcasts:Burn Wild

Burn Wild sets out to answer the question "How far is too far to go to save the planet?" and I nearly gave up halfway through as it seemed to be giving too much credence to the idea that as long as nobody is hurt, it's OK to blow shit up. In the end though it doesn't let anyone off the hook, and several people come to realize that they did no good and may even have done some harm. Worthwhile.

The Missing Cryptoqueen

Review:Podcasts:The Missing Cryptoqueen

I'd never heard about Dr Ruja Ignatova or OneCoin before listening to this podcast from Jamie Bartlett that makes the case that it's a ponzi scheme rather than a revolutionary crypto currency. She has been in the news very recently and so I'm expecting another episode or two soon.

Although, to be fair, crypto has always bothered me as a way to describe Bitcoin and other blockchain technologies. Crypto means hidden or secret. Blockchains by their nature are very open and not secret - that's kind of the whole point. A ledger that is not distributed is a database, like what a bank uses (or a ponzi scheme). So maybe she has a point.

TV

Black Summer Season 1

Review:TV:Black Summer Season 1

My favorite part of any apocalypse drama are the moments when civilization is just starting to fall apart. Like some anti-Gibson, the end of the world is here, it's just not evenly distributed. Black Summer just goes straight to the biting. No foreplay.

Why make yet another zombie show? Couldn't it be Mormans trying to fill the book of the dead a bit faster? Or environmentalists trying to slim the population down by seven billion people and you get infected by podcasts? Netflix, call me.

I loved early Walking Dead, and I think it was mostly the dialog. Those scenes where the cast is strolling through Georgia, on the way to some salvation that won't pan out and shooting the shit along the way. Black Summer probably has five lines of dialog in total. As if it realizes it has nothing to say the runtime reduces with each episode - 44 mins at the start, halved to 22 mins by the finale.

It's tense though so I'll probably still watch season 2.

Dark Summer Season 2

I tried. It's now winter (get it, the sequel to summer) and there seems to be a completely different cast. I wasn't that invested in the original lot, but I certainly don't even begin to care about these people. It keeps everything that made season 1 miserable but drops the tension. I made it to halfway through the second episode and then bailed.

(All images included with ITHCWY reviews are the property of their respective owners and are used to illustrate reviews only.)

Add your comment...

Related Posts

(All Reviews)

(Published to the Fediverse as: Reviews for February 2023 #reviews #dmsinclair Reviews for February 2023 )

Catfood.Shapefile 2.00

Updated on Sunday, March 12, 2023

I just released Catfood.Shapefile 2.00, my .NET parser for ESRI Shapefiles.

The big change is that I have migrated to .NET Standard 2.0. This makes it possible to use from .NET Core as well as classic .NET Framework from 4.6.1 up. If you need to use an older version of .NET Framework then you'll want to stick with Catfood.Shapefile 1.60.

Catfood.Shapefile is now also available via NuGet. This is the recommended way to install. The source code is still available on GitHub.

Add your comment...

Related Posts

(All Code Posts)

Upgrading from word2vec to OpenAI

[1536]

In 2018 I upgraded the related posts functionality on this blog to use word2vec. This was hacked together by averaging the vectors for interesting words in each post together and then looking for the closest vectors. It worked quite well, but the state of the art has moved on just a little bit since then.

OpenAI has an embeddings API and recently released a cheaper model called text-embedding-ada-002. The vectors have 1,536 dimensions, a pretty significant increase from the 300 I was using with word2vec. Creating vectors for all my posts took a few minutes and cost $0.11 which is pretty affordable. As you'd expect those related posts are now significantly more related and useful. Thanks OpenAI!

I shared some code previously for the word2vec hack. This is a lot more straightforward - call the API with the post text and then compare the vectors with cosine distance to find the most related. It works well for search too.

Add your comment...

Related Posts

(All Code Posts)

(Published to the Fediverse as: Upgrading from word2vec to OpenAI #code #ml #openai #ithcwy #word2vec Using the Open AI embeddings API to find better related posts for a blog. )

Migrating a C# Integration from GA3 to GA4

Updated on Saturday, May 6, 2023

GA4

This blog has a couple of Google Analytics integrations - the popular posts list is pulled from GA, and the unnecessarily accurate count of non visitors in the footer. I just migrated from the GA3 API to the GA4 API. The backend for this blog is ASP.NET MVC with .NET 4.8. One day I might catch up with the cool kids and try to get on .NET Core, but not today.

Here's where I stubbed my toe:

I'm following this code sample to make my first GA4 call. After installing the NuGet package I couldn't find BetaAnalyticsDataClient anywhere. It turns out that there is a Google.Apis.AnalyticsData.v1beta package and a Google.Analytics.Data.V1Beta which is only available if you check 'Include prerelease' when searching. You want the second one. I'm not in love with BetaAnalyticsDataClient as a class name, it suggests all sorts of breaking changes are coming. My GA3 integration has ticked over for years with no changes. Maybe GA4 is going to be more like Google Ads and shank you with breaking changes every few months. Moving on...

Wow the error messages are good. Kudos to the API team. I'm so used to cryptic bullshit but this API tells you what you're doing wrong and sends back helpful pointers and even URLs. Every API should be this friendly. I got through the remaining problems fairly quickly because of this.

The code sample passes the property ID as 'property/nnnnnnn' but the API is expecting 'properties/nnnnnnn'.

I'd been using a ServiceAccountCredential created from a .p12 file for GA3. This doesn't seem to be supported for BetaAnalyticsDataClient but I was able to generate a new credential with a .json serialization of the credentials and passing this to BetaAnalyticsDataClient worked fine. I had a permission denied error, this was because I hadn't added the service account email address to the property and doing so got me some data.

The client library is pretty classy (as in too many classes). Creating a filter to exclude internal users involves four nested classes - a FilterExpression that has another FilterExpression for a not condition and then this needs a Filter and the Filter needs a StringFilter. Tedious. And including enums for metrics and dimensions is too much trouble so adding those now requires Metric and Dimension classes but these are just initialized with a string. The list is here.

Lastly when it comes to running the thing the site won't start and says:

"CS0012: The type 'System.Object' is defined in an assembly that is not referenced. You must add a reference to assembly 'netstandard, Version=2.0.0.0, Culture=neutral, PublicKeyToken=cc7b13ffcd2ddd51'."

Presumably due to some NuGet horror or other. Adding that reference indeed fixes the problem and hopefully doesn't create a new one.

I am now technically if not emotionally prepared for GA3 to be switched off.

Add your comment...

Related Posts

(All Code Posts)

(Published to the Fediverse as: Migrating a C# Integration from GA3 to GA4 #code #ga4 #ithcwy #c# Some tips on avoiding pitfalls when migrating a C# application from Google Analytics 3 to Google Analytics 4. )

Timelapse of San Francisco Clouds After Various Atmospheric Rivers

Timelapse of San Francisco Clouds After Various Atmospheric Rivers

Four time lapse sequences of cool clouds over San Francisco at sunset following the January 2022 sequence of atmospheric rivers.

Add your comment...

Related Posts

(More Timelapses)

(Published to the Fediverse as: Timelapse of San Francisco Clouds After Various Atmospheric Rivers #timelapse #video #clouds #sanfrancisco #sunset #4k Post atmospheric river (January 2022) sunset time lapse from San Francisco, California. )

Server Migration Complete

2022

ITHCWY just migrated to a new server with more capacity. From clouds to pollution to covid monitoring there is a lot going on under the hood here these days. I can already see that search is a lot more snappy.

Man, moving Windows servers just sucks. Internal GDI error - that's a permissions issue obviously (not my first time). Can't publish? You must need to install Web Deploy twice and then uninstall it and install it one more time for luck. Web.config error at startup? Of course Windows Server 2022 still doesn't come with URL Rewriting and you need to install it from some corner of the MS website like an animal. Yes, I should use something else but my custom CMS is probably reaching the same level of complexity as an F16 and I'm just one person. Enjoy the fast search!

Add your comment...

Related Posts

(All Etc Posts)

(Published to the Fediverse as: Server Migration Complete #etc #ithcwy #microsoft ITCHWY has migrated to a beefer box and a new OS and still seems to be working. Oh, and search is faster! )