Two years ago, I launched a start-up to "Uberise" the cleaning, handyman, electrical and air-conditioning service industry in Singapore. I was awestruck by the power of the Internet in connecting consumers to providers, and motivated by the promise that platforms like Uber could empower tens of thousands of freelance "micro-entrepreneurs" to improve their livelihoods through these platforms' "asset-light" models.

I had just finished graduate studies in Oxford University after my degree in London, and had interned at a consulting firm. But I wanted to do something different. Riding on the belief that "software will eat the world", and the vision that it could help build a more egalitarian society, I became part of a broader phenomenon that has become known as the "sharing economy". Poster-child tech start-ups like Uber and Airbnb are pioneers of this trend, with spin-offs in numerous industries, for instance in groceries (Honestbee), pet-sitting (Gudog) and food delivery (Foodpanda).

Unfortunately, that vision is little more than an unfulfilled promise.

I am sceptical that the sharing economy creates a more connected and egalitarian society. In particular, I argue that its business model is an unsustainable alternative to mainstream commerce.

SHARING ECONOMY V1.0: AN UNSUSTAINABLE WINNER-TAKES-ALL MODEL

Companies like Uber have risen meteorically through unprecedented rounds of venture funding. While this cannot be understated, it is also becoming clear that such sharing-economy models are unable to turn a profit.

A CNN report from 2016 is telling - Uber burned through more than US$2 billion (S$2.7 billion) in 2015, while previously losing US$1 billion every year in China.

The "asset-light" model may seem ideal because the marginal cost of providing the product is near zero. It is often said that such companies own few physical assets, and while an external party provides the service, the platform takes a cut from each transaction. The idea is appealing, but also grossly underestimates the hefty expenses that go to pay Silicon Valley developers, engineers, data scientists and marketers in the development and running of the platform. It is superficial to judge the profitability of a start-up based on it being asset-light alone.

In other words, the cost of developing the tech platform is understated. It is true that once set up, the additional or marginal cost of adding one more user is close to zero. But one can't ignore development and marketing costs.

On-demand companies are also starting to lose the legal battle over how their service providers are classified. Platforms like Uber and Helpling.com classify their freelancers (cleaners, drivers or helpers) as "independent contractors", relieving them from contributing to their workers' social security accounts (Central Provident Fund, in Singapore's case), insurance premiums and medical expenses.

Job risks are pushed entirely onto the sub-contractors, who join these platforms with the "promise" of attaining a higher and more stable income stream.

It is clear that the skirting of labour laws on these semantic technicalities is backfiring. Quite recently, Homejoy - a San Francisco-based platform for on-demand cleaners operating in America and Europe - was shut down due to lawsuits.

In a recent British employment tribunal, Uber came out on the losing side of a landmark ruling denying them the right to classify their British drivers as "self-employed", to protect the rights of these workers.

These legal challenges are a clear signal that sharing-economy platforms running afoul of regulations will become costlier, and eventually even illegal. The dependence of micro-entrepreneurs on these forms of insecure employment gives rise to the threat of chaos and uncertainty in the job market.

FAILED PROMISE TO EMPOWER WORKERS

Worse, many of these platforms generate misleading promises related to compensation.

Helpling.com claims that its cleaners earn about $16 per hour or up to $1,500 a month if they are hard-working enough.

But this overlooks many things: the cleaners' 1 1/2 -hour average travelling time between jobs, their transportation fares and the cleaning supplies they purchase.

Also, if an appointment gets cancelled, there's no loss to Helpling; it doesn't have to pay the freelancer. If the freelancer is taken ill, that is his problem. I discovered that painful reality when I worked as a cleaner some time ago.

One might argue that the gig economy is meant to supplement income, not to replace actual jobs. But the notion of working on one's "own time" is also an illusion.

As these platforms start to become more powerful, they begin to exert more control over their freelancers by imposing conditions related to accepting jobs and job performance. In platforms like Uber, freelancers can be kicked off the platform if they become picky about selecting work.

Ultimately, these sharing-economy platforms are making big money for their investors by creating riskier and more precarious forms of low-paid work for those who provide services on their behalf.

SHARING ECONOMY V2.0

As it stands, it is clear many of these sharing-economy platforms cannot be trusted to guarantee a safe customer experience or create fair and dignified employment for their community.

Nonetheless, there is hope.

There is a second wave of sharing-economy start-ups that are abandoning their dependence on independent contractors in favour of well-paid, full-time employees.

These platforms harness the good bits of technology developed over the last few years - ease of booking, peer-reviewed systems and digital records of all past transactions - while operationalising the importance of compensating employees well through a "good jobs strategy" movement popularised by Massachusetts Institute of Technology Professor Zeynep Ton .

In the United States, companies like Eden (tech support) and Hello Alfred (butler service) are hiring full-time employees instead of freelancers. Closer to home, start-ups like Dah Makan, a Malaysian food-delivery platform owning its entire operations in-house, do likewise.

At Nimbusforwork.com, our platform aims to disrupt the office-management space with our full-time office cleaners and labour support for things like handyman work, office cleaning, event cleanups and office supplies. We want to marry powerful marketplace technology with exceptional service.

The common thread between these next-generation start-ups is the emphasis on putting people before algorithms.

The bet here is investing in a happier workforce to increase customer satisfaction and retention while taking full responsibility for service standards and employee churn.

Whether this shift in the sharing economy provides a more sustainable and ethical business model remains to be seen. What is clear, however, is that the old start-up ways of valuation-chasing and skirting regulations, while eroding the rights of freelance workers, need to change.

• Daniel Thong is co-founder and chief executive of Nimbusforwork.com, an office-management platform offering cleaning, office supplies and handyman services.