AI助力公关……是吗？

Costa工作：独立软件工程咨询公司的启动广告案例

最终，一切都准备就绪，世界上效率最高的软件工程咨询公司ERP: Costa Works即将启动。

因此，我做了一些视频来做演讲的热身，并且调整了我的视频制作流程；现在我不再害怕摄像头(或多个)，并且基于montage-as-code脚本(FFmpRb DSL)可靠地从“录制设备云存储”(如GoPro++)到“视频社交网络”(如YouTube等)生产媒体.

那是为下一级个人IT咨询专业化做准备的最后一步；也许您不会感到惊讶，我第一份工作是一个个人合同工作（它也涉及到年轻的孩子和机器人！），那是当时未成年雇佣的标准做法，但我的第二份工作也是合同工作，然后我再次返回到合同工作——从我的角度来看是最有效率的——一次又一次。每一次都有所改善（我非常喜欢我在Astrails咨询公司的时间），我获得了一些知识和经验，但现在，我还预先集成了流行（实用）技术在一种超级机器人工具箱中，它可以启动任何项目知识产权（价值）的即时增长！不，说真的，这些组件需要几周时间来开发，可以服务多年，这就是我在项目中所做的，投射出我的有效方法，这甚至比技术工具箱更重要。

下一步是跳出（“领英”）泡沫，以一种可能吸引不认识我的正确人群描述我的服务。我认为一个约5分钟的剧本介绍视频应该就可以解决问题。于是我写了这个剧本，经过一些初步批评和编辑后，看起来是这样的：

#hl0 - coming in
Hi, my name is Costa Shapiro, and this is another video by "costa works".

#hl1

"costa works" is looking for new commercial projects now, so here's a shameless
self-promotion.

#hl2

Are you an IT entrepreneur or project manager caring about your team's
efficiency?
Do you struggle to understand the value of work your engineers prioritise
-- every planning meeting?
Would you rather like to
(1) set capability goals,
(2) get price and specs,
(3) plan and execute?

#hl0 + coming in
[And this is where I come in.]

As a management team member, I can help separate and maintain the processes
of knowhow research, product definition, and "pragmatic progressive" development.
To note, it's essential these business processes are separated in an IT project,
with different people responsible for each, preferably.
The "pragmatic progressive" development process -- which I'm ready to establish
and lead for your project -- has the Pareto principle at its core and this means,
["work%|results%" === "20|80"
FYI, 0.2 + (1 - 0.2)* 0.2 / 0.8 + (1 - 0.8)* 0.8 === "36|96"]
as a project manager,
- first, you never start with nothing, there's no system "ramp-up":
  - your _own_ minimal (wrt skill requirements and other dependencies) project
    infrastructure -- both for development and for production -- is all ready
  - non-principal component and interface technologies are integrated already,
    so you can focus on increasing _value_ and not on trivial problems
  - major business scenarios can be run very early with the production system;
    having dedicated UI (instead of integrated generic interfaces) is _optional_
- then, you're always _optimising_ the system, i.e. improving its support
  of essential scenarios involving users, operators or developers
  - through reviewing research results, planning and executing modular tasks
    -- after reasonable "resource estimation against benefit prediction" analysis
  - while another part of this development process is ongoing formal (testable)
    system specification -- of those scenarios, following BDD principles
    -- which also certifies the validity of these optimisations and the actual
    system's value
[

start with most engineers:
start with nothing of your own

plan and prepare for everything ahead

implement actual business logic

fix bugs

finish with something of vague capabilities and value

|vs|

start with costa works:
start with many abstractions implemented in _your_ system,
so many business processes can start right away

define beneficial ways to optimise system

get resource estimation and task specs

implement profitable optimisations

finish with sum of your valued optimisations exactly

]

#hl2
I understand business momentum as well as I understand engineering
procrastination -- which is all the organisation and preparation for work,
and such "necessary work" as implementing exactly the same functionality
like at their previous project, "just better", bringing zero value
into the system at the beginning of its development.

For a business -- especially in its beginning, it is essential to be
"usage-first, structure-later", or for an IT business, "data before code",
to execute service-feedback processes as early as possible,
and since a business rarely has a steady start-up, development resources
-- just like computing resources -- should be dynamically scalable as well.
[you can mark features Beta, if you like to, but don't put cart before the horse]

Scaling computing resources is largely a solved problem in the industry,
they also say "scaling is a good problem", so, you are unlikely to fail at that
either with or without my expertise in your project.
[Scaling   Computing   Resources]
However, what I called "scaling development resources" means engaging
and dismissing developers of particular expertise within a project as well as
changing component technologies or other dependencies which induce the expertise
needed for system maintenance.
The secret to doing this kind of scaling efficiently is, of course, following
the IT developer community, choosing the technologies as appropriate,
and then, following the community standards sensibly, while keeping
the "business IP core" as small and powerful as possible.
[Scaling  Development  Resources: (the secret)
- dev community
- choosing tech
- community std
- business core                               ]

For example, yes, you can probably _somewhat_ optimise your system usage
with a branded graphical user interface (which you will need React or Flutter
or Next developers for), or, for the time being, it can be just an AI-aided
chatbot in a standard customer's messaging app (which you will need some other
expertise for, of course).
[not even trying to guess the leading chatbot framework at the time of viewing]

#hl1
Unfortunately, the development resource scaling has no one-fits-all solution
_just yet_ -- until engineering-manager-GPT is available -- luckily, there are
system professionals like me who can help you manage your high-tech development.

My individual digital information environment vision (including its many
interfaces, abstractions and resources) -- for developers, operators and users --
is a result of the life-long interest and specialisation in software engineering.
[e.g. an M.Sc. in SW Eng. from Technion CS]

So, from my academic background, through my end-to-end ownership experience,
to my pragmatic professional tech toolbox, "costa works" can probably work
for you.

[final slide with QR]

然后，遵循多个朋友的建议和自己的好奇心，我第一次尝试使用ChatGPT（“OpenAI”）来实现真正的（商业）目标。很抱歉说，但我真的被大大地失望了，然而，我并不感到惊讶，因为它现在引起了广大公众的关注，而这些人在想象力上灾难性地缺乏；对于更有想象力的人，他们也是近几十年来的技术迷，在代理人通讯风格方面并不新鲜，在其智能能力方面则…

如果你的生意只是向企业午餐派送三明治，我相信机器人的上下文深度足以帮助这种交流，以协助交互时不够善于口语的员工。

所以，如果经过训练的模型具有许多变体和关联，这必须是一个基本概念（也是情感上与所有人都非常亲近的），对于大多数人来说，这个机器人可以像更原始的游戏中的人一样令人印象深刻 年轻人具有很多想象力。但是，如果您试图传达甚至稍微更复杂的想法，您只是在您扭曲的“含义”周围获得“起沫”：“语法”结构缺少知识性信息的随机改词（和改写） - 有时是有损的 - 原始文本。

我不确定我尝试与机器人“聊天”的经历是否足够有趣（顺便说一句，“聊天”感觉像经典的模糊逻辑查询[另外，结果的获取、链接和无关紧要的内容]，因此在这方面没有什么新的东西），但如果没有示例，这篇文章就不完整，所以这里有一个会话日志：