Are you a big AI business vendor making terrible AI business decisions? We can help

Congratulations! As CEO of a giant tech company, head of a sovereign wealth fund, or a VC bored with megayacht leapfrog, you have billions of dollars of other people's cash to spend. You want to make a difference. You want to be a success. Most of all, you want to look good doing it. Right now, nothing looks better than AI to …

[Anonymous Coward]

"What are you buying - and when you’ve bought it, what are you selling?"

Possibly the best bit of advice I have seen in a very long time.

"Align your core mission with bankruptcy." Golden!

The smart strategy if you can go fully private is to let the others go first, hoover up their clever former employees displaced by AI, advertise an AI Enhanced code generation (or whatever) service at a premium but apart from some front end eye candy use your newly acquired human talent to put in the hard graft.

In case anyone thinks that this is false or misleading - the service has been manifestly enhanced by the uncritical adoption of AI... albeit by the opposition.

Once the AI tulips wither and die in the frosts of the coming winter you can cast aside the mantle of the AI pretense and acquire any worthwhile assets from your now defunct competition at fire sale prices.

[Jedit]

"Align your core mission with bankruptcy."

It's not like that's a joke. Startups with 100% external investment can waste as much of their backers' money as they want, because there's one part of the funding that will never be flushed down the toilet: their salaries. They can promise anything, and even try to deliver it if they think they can. But if they don't deliver and the company goes under, they still got paid.

[sketharaman]

Re: "Align your core mission with bankruptcy."

Well said. As I keep saying, only 20% of startups succeed but 100% of startups provide employment and pay salary. IMO this 5:1 ratio between fruits of unearned success and real success is the real reason why everybody incuding politicians keep talking up startups.

[andy the pessimist]

options/fallbacks

If you are spending a fortune on AI.

Can your data centre be designed so it can be a regular data centre or an AI data centre? Some loss of performance is acceptable. Or is it one huge spin of the device.

Can the infrastructure be organised to cover both options?

Is there a proof of concept AI application doing similar processing?

Can somebody debug the software.

Are you feeling lucky.

[Doctor Syntax]

Re: options/fallbacks

"Are you feeling lucky."

Feeling lucky and being lucky are two different things.

[breakfast]

Using AI for a competitive advantage

In the current environment it is easy for a tech company to win with AI. First get a few teams in, run a few POCs, report on how well they went.

Now report on how you are using AI more widely across your company. Report that it is saving 30-50% of developer time, that your productivity has increased by 80% while reliability is 35% higher than before. This will astonish and horrify your rivals, who will have to jump on board the AI train to keep up, they will be desperately trying to figure out what you did and how you achieved those numbers. Their board will be so impressed with what AI is capable of that they start firing their more AI-sceptical engineers who are inhibiting the opportunity to integrate AI into every part of their work flow.

All of this is the standard pattern we have seen across the sector. The trick is that you're not using AI internally to any significant extent. Instead you keep using real human developers and maybe even recruit some of the AI-sceptical engineers your rivals are getting rid of. Even if your code is written more slowly it is going to be more reliable, secure, and, as you won't be forcing skilled developers to babysit the toddler-level outputs of LLMs, both quicker and more pleasant to work on. Meanwhile your rivals will continue to tie themselves in knots desperate to match your completely fictional numbers as they rapidly vanish in your rear-view mirror...

[b0llchit]

CEOs don't fail...

...investing in an AI failure could look bad, very bad, or truly awful.

As professional CEO, I can not blame myself for any failure. It is my very great and true experience, as CEO, that, in case of any failure, my underlings always have failed to execute my orders correctly. Therefore, I, the CEO, can never be blamed for any failure. It is always and truly the fault of everybody else when failure strikes.

[Decay]

Re: CEOs don't fail...

And when you write your vision statement and strategy paper ensure it's open ended enough to be interpreted anyway you need if and when it goes tits up.

Something like

Strategic Vision Framework: F.U.C.T.

Flexible

Unified

Customer-centric

Transformation

“With F.U.C.T., we’re charting a bold, adaptable course for the future—because if we don’t stay ahead, we risk getting left behind.”

Optional tagline:

“If you’re not aligned, you’re F.U.C.T.”

[joeldillon]

I would be um somewhat surprised if Windows 3.1 let alone 95 was 'largely' assembler. DOS, maybe.

[disillusioned fanboi]

Perhaps for "lift-and-shift"?

I work in contact with zOS Mainframes. There is a trend to switch to Intel Linux, and many companies sell services for that.

I've seen projects where applications on Linux have to produce the exact same output as the application on zOS.

So in that use case I could indeed see an AI coder doing a large part of the conversion process.

For example if converting from COBOL to Java, you could have a team of Java coders, and the AI reading the COBOL and producing a first draft of the Java code.