Will AI and Space Breakthroughs Continue in 2023

The biggest developments in AI were the Openao Dall-e 2 and ChatGPT. OpenAI said Dall-E 2 was generating over 2 million images a day.

People are using ChatGPT to write stories and articles and then having Dall-E 2 create the pictures for the ChatGPT article or story.


SpaceX had 61 launches in 2022 up from 31 in 2021 and should have over 100 in 2023.

We are still waiting for the first launch of the Super Heavy Starship to orbit.

SpaceX is launching Gen 2 Starlink satellites to orbit and should be launching them once per week. This could enable 600 orbit to unmodified cellphone text and voice communication. This could support tens of millions of customers by the end of 2023. If the Super Heavy Starship is successful and starts frequently launching payloads then the Gen 2 deployment will speed up by four times launch for launch. If the Starship gets to a higher launch rate then the speed of deployment will increase. A fully deployed Gen 2 with 30,000 satellites with 30 Mbps direct to cellphone communication will enable 2G 10 kbps communication. It will be more difficult to get constant service to dense cities. A dozen gen 2 STarlink satellites in view of a city would be able to provide 300,000 simultaneous 2G connections (not fully utilized) to a city. However, 1 kbps connections for about 3 million people. It would be a decent mix of 2G and 1G communication to billions of people.

This does not even mention all of the higher bandwidth Starlink to satellite dish communication. This would scale more limited by the production rate of satellite dishes. We are at about 150k per month dish production. SpaceX can learn from Tesla to reach millions per month in manufacturing.

Self Driving and Driving Automation

Baidu on Friday was granted the first license to test driverless vehicles on roads in Beijing, and would add another 200 robotaxis to its network across China in 2023.

Tesla made progress with FSD. It still has widely released it in North America to 285,000 customers.

Cybertruck and Semi

Tesla introduced a few dozen Semi in 2022. They plan to get to 50,000 in 2024. 2023 is the year of scaling. They have the 40 GWH of batteries from Nevada. There are billions in grants and subsidies especially in California and New York. There are massive cost advantages for the Tesla Semi over any other electric Semi. The Tesla Semi is more cost effective than diesel trucks.

The Cybertruck should be released in them middle of 2023. It should have adapted a lot of the Semi truck. This should enable the Cybertruck to have superior towing and cargo capacity not only against electric pickups like Rivian or the Ford Lightning but also the gas Ford F150. The Ford F150 has a 23 gallon tank and it goes from 20 mpg to 9 mpg under heavy load which is 460 mile unloaded range to 207 miles loaded range. A Cybertruck with Semi 1000 volt and better engines should go from 500 mile range and 250+ mile loaded range.

Aging Reversal and Antiaging

Quantum Computers

Fusion and Advanced Nuclear Fission

In December 2022, General Fusion has demonstrated plasma energy confinement times, plasma temperatures, and compression system performance that support meeting its goal of 10 keV (100 million degrees Celsius) in the company’s integrated Magnetized Target Fusion (MTF) demonstration, being built to 70 percent scale of a commercial machine at the UK Atomic Energy Authority’s Culham Campus.

Creating fusion energy with (Magnetized Target Fusion) MTF starts with a hydrogen plasma injected into a vessel lined with liquid metal. Next, high-powered pistons compress the liquid metal around the plasma, squeezing and building pressure until the plasma reaches more than 100 million degrees Celsius – and fusion occurs. The technology depends on three core factors to be successful: hot plasmas that hold their energy long enough to be compressed, a synched compression system, and a stable fusion process that will increase neutron yields and temperatures when plasmas are compressed.

The plasma injector exceeds requirements with 10-millisecond self-sustaining energy confinement time.

About 24 new nuclear reactors should be started in 2023 or 2024.

7 thoughts on “Will AI and Space Breakthroughs Continue in 2023”

  1. The progress doesn’t just keep coming, it accelerates, as a greater wealth of knowledge fuels the incentives and capabilities to achieve more.

    Something someone else alluded to: major changes do not happen in a heart beat. You cannot, for example, assert that the technological singularity that was, say, agriculture, occurred on a specific day in a certain year (as you can with the meteor strike that killed the dinosaurs).

    We look back now and clearly recognize the industrial revolution, but not on a singular day. It is only looking back from a distance that we get a rough idea of when it became ubiquitous and effectively unstoppable.

    It’s certainly not the moment a new discovery is made. Philosophical, ethical, religious, political, and economic concerns can easily delay things. This was especially true with the printing press, for example.

    The internet existed before 1993, but if we want to pick a pivot point where it really took off, that might be it. We certainly could not see it at the time, or what would follow, not in great detail.

    I’m thinking that, looking back from far enough in the future we will mark Cognitive Automation, devices taking over tasks that require routine thought processes, as being right around where we are now, maybe a few year from now, or perhaps a short time ago. Somewhere around thirty years after the internet, which followed electronics by about sixty years, which followed the industrial revolution by about 120 years, which followed printing by about 240 years, and so on.

    So sometime in the 2040s for the next, unless it is a double singularity, which is certainly possible, perhaps synthetic intelligence (not just a work around for intelligence, like AI is), and radical life extension.

    But then you get another about 7 years later, and another 3 years following that, and, finally, you get a year (perhaps 2053) where they never stop. This is obviously ridiculous.

    But by all means, read Vernor Vinge’s classic “Marooned in Realtime,” to get some potential hints as to what this might be like.

  2. Oh boy, are we for a treat in 2023.

    SDCs are well stablished as a developinga area. Fusion made a splash, but it’s really an incremental step. Of all the innovations, ChatGPT had the biggest impact on the public’s imagination. Both good and bad things.

    And it’s understandable. The complexity and sophistication of the GPT3 LML is a new thing for most people, and makes some believe they are dealing with a sentient entity. I could myself get that perception while working with it producing small programs in several languages.

    You can ask it for editions on the data structures and flows in fully natural language, and it obliges nearly flawlessly, as long as you use the right specification language.

    You can feel it understands the abstract entities it is manipulating, and you can have discoursive approach with it, modifying parts of the existing solution and getting increasingly big programs. It is limited in the size of the single replies it can produce, but it can continue adding more if you prompt it.

    If GPT3 could do this, the upcoming GPT4 will freak us out big time.

    • I’d consider the fusion “breakthrough” more of a proof of concept – up til now they hadn’t proven that it was really possible to get ‘ignition’. Before, they only had theory.

Comments are closed.