Will AI end with SQL, Is this the end of SQL?

  • Comments posted to this topic are about the item Will AI end with SQL, Is this the end of SQL?

  • The concept of AI in the picture near the end of the article is actually pretty grotesque, IMHO.  Sex doll for both the body and the mind.  Possibly the worst drug ever. 😉

    As for the rest of the article, it's very well written but it's about a very simple subject.  I can certainly see it being convenient to not have to figure out how to import a CSV but that was actually done even before the advent of SQL Server.  I have similar comments about the other sections and so won't actually bore anyone with such comments.

    And, yes... I know AI is still in it's infancy... it could do some incredible things  and allow humans to do some incredible things, as well...   and I'm very concerned about multiple aspects of it.

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • Maybe, but not for for a while yet. The data sample used to illustrate this article is extremely simple, compared to that which most data migration specialists come into contact with every day.

    Where a db application doesn't match the evolving business needs of the client, the client often applies... 'inventive' techniques to solve their problems. These are frequently undocumented evolve (still undocumented) over time. At the point of migration to a new system, the unravelling of these issues does not appear (to me at least, for now) to be a strength of AI.

    And just as a point of interest, see this LinkedIn article

  • I agree with roger-562742.

    Most AI examples I see that touch programming so far are pretty basic.

    Also, while copilot in Visual Studio Code can make an impression by suggesting 15 lines of code, it is often wrong.

    Would anybody in his/her right mind let AI write anything important?

    For a good laugh, read this about rot-13 "encryption" and chatgpt:

    https://jameswillia.ms/posts/chatgpt-rot13.html

     

  • The AI suggests that the lower cost of flights in November might deter discretionary travel. I'd have thought that lowering costs would encourage it.

    --
    Scott

  • This part:

    "Cost - With fewer holidays and events, November may have a lower demand for flights which could drive prices down. The lower cost of flights could deter some discretionary travel."

    Just doesn't make sense - if anything, driving prices down would tend to favour discretionary travel, not deter it.

    Other parts of the Claude response are hardly succinct - the end user has to trawl through spurious gumpf to actually find the actual answer to the question.

    Maybe one day - but based on this, not very soon

  • The world is definately changing from a technology point of view. How we work will definately be affected. But I do still think, data security governance and warehousing as a whole will still always be pivotal. I can't imagine a world without all those tools above mentioned. Especially Microsofts highly anticipated ecosystem ADF. I don't think that will change in atleast the next 5 years it will grow and AI will be plugged in on top of all these tools. Because you still need to orchstrate data carefully especially with LLM's being the the front runners this year and years to come. You will always need code no matter if the horrible LLM is wrtiting it or the good human. Making sure that garbage code is not used will make it very hard to say that. Any business person to write code or ask the LLM to generate code for stats. One line of inacurate line of code or whatever can skew figures and I think reporting on wrong stats is worse than letting the capable body go. Just to cut cost. Here is my take i think AI will definately make all dev's code experiences a little bit easier like all the other older languages that was difficult to understand and write became easier and easier to the point where its going to probably be natural language 75% in the next few years. I simply don't believe Devs or DBA's will fail to exist.

  • I don't think so AI will going to replace BI and DBAs job especially with enterprise client because of complicated scenarios and requirement. For example, data governance, BI framework and etc.  However, it will make our job easier and faster.

  • So far, AI seems to be capable of complex procedural tasks. It is like solving equations manually versus using MATLAB. Now everybody can solve linear equations, by means of MATLAB or AI (I am sure such AI exists that solves liner equations). However, no AI will ever put together set of equations to solve actual real life problem, with real life constraints, same goes for ATLAB and like. Even some more known riddles can (or will be possible) to solve bay AI. For example, "A brick's weight is a a pound and 1/2 of the brick. How much is the brick's weight?". One can certainly teach AI that the answer is 2 pounds. Riddle translated in mathematics is this:

    (Brick Weight)= 1 Pound + 1/2 (Brick Wight) which is the same as 1/2(Brick Weight) = 1 pound, which is the same as (Brick Weight0 = 2 pounds. I am sure AI will be able to solve x = 1 + (1/2) * x  in a jiffy. However, the problem is how to come up with the equation.

    I am sure AI can be trained to solve anything, even to look like logical reasoning to untrained eye.  AI to create something? Not so sure.

    No worries, DBA and database designers cannot be replaced by AI. Although, there will be attempts,  many people will take it hook, line ad sinker. Whoever teaches AI to design databases based on Microsoft templates is in for a big surprise.

    As consequence of more and more people relying on AI will be disastrous - less and less people will be able to use NI - natural intelligence. Or it could be poor translation. In English, intelligence means two things - 1) being clever, and 2) being well informed, having lots of facts in one's head. More and more people are forgetting about 1. Unfortunately, in time there will be less and less people capable of understanding difference between being clever, and having crammed lots of information, true or false.

    Zidar's Theorem: The best code is no code at all...

  • This currently works well for simple datasets (csv file, excel file, a table, etc.) but anything more than a single table select statement and things get too complex for it.

    AI is a great resource for people working with databases. It will be exciting to keep up with the changes and to think of ways to incorporate AI into our daily routine.

  • Introducing Copilot in Azure SQL Database (Private Preview)

    Of course, of course ready or not here it comes.  As others have noted AI is quite useful for ginning up boilerplate code in certain circumstances.  Afaik this copilot is built into the Azure Portal SQL Editor which I've never used.  Coming soon (I'm sure) to SSMS and Azure Data Studio

    The long term effect is hopefully more data models get built and deployed

    Aus dem Paradies, das Cantor uns geschaffen, soll uns niemand vertreiben können

Viewing 11 posts - 1 through 10 (of 10 total)

You must be logged in to reply to this topic. Login to reply