But OpenAIs rapid ascentit launched in late 2015 and shared ChatGPT with the public two months agowill no doubt leave some entrepreneurs wondering if theyre doing something wrong with their own less-noticed ventures. 

Elad Gil, a widely respected Silicon Valley angel investorhe made early bets on Airbnb, Instacart, and Squarebelieves the fact that ChatGPT is down all the time right now is a great sign of product-market fit. Its because too many people are using it. Thats a great problem to have.

Gil made the comments on an episode of the Logan Bartlett Show podcast Friday. The alum of Google and Twitter noted the problem OpenAI faces after Bartlett, a software investor at VC firm Redpoint, asked about his thinking on product-market fit when considering investments.

One sign he looks for is positive testimonials from customers and users, he said: It just really comes out in terms of the enthusiasm that small initial cohort has for the product.

But also, he added, If a product is broken all the time but everybody keeps using it, theres clearly product-market fit, noting he witnessed that in Twitters early days and sees it now with ChatGPT.

On the flip side, he said, many ideas will simply not take off, no matter how much time an entrepreneur sinks into them. He said the lore in Silicon Valley that you should grind eternally and then eventually something will work is wrong, noting people have wasted years of their lives because of such bad advice.

People end up spending years and years and years of life just grinding away on something that isnt going to work, because maybe it will work if I do these three more tweaks, and maybe it will work this month if I keep going, he said. For a very small number of cases that happens, but for the majority it works immediately, or near immediately. 

While its true entrepreneurs might need to power through tough times during a recession, he added, When times are good, the worst advice you can give somebody is keep going no matter what. 

Theres enormous opportunity cost on your time, and most things dont work, Gil said. Most of the time, you should actually figure out when do you give up and when should you actually quit. Its really hard to know. 

Meanwhile, when an idea works, it tends to work very quickly, something hes seen repeatedly with companies hes worked at and invested in over the yearsand now sees with OpenAI and ChatGPT.

The reality is that most of the companies, not all, but the vast majority of companies Ive been involved with that worked, ended up working pretty early. And once they started working, they just kept working.

Hes also watched as others have made the same realization.

One thing Ive noticed is that people who have worked on things without product-market fitthat they thought had product-market fitwhen they finally go and work on something that truly is working, they realize the immense difference and the degree to which they were fooling themselves.

In the first case, youre chasing everybody and every sale is gruesome and everything is a grind, he said, but with the latter, its a case of, Hey, people keep calling me. And so its this transition, and until that happens you dont realize what that really feels like.

Learn how to navigate and strengthen trust in your business with The Trust Factor, a weekly newsletter examining what leaders need to succeed. Sign up here.

At Bankman-Frieds arraignment in early January, Assistant U.S. Attorney Danielle Sassoon accused the FTX founder of transferring Alameda assets from a crypto wallet, successfully imploring the judge overseeing the case to restrict Bankman-Frieds access to company funds and implying that his word should not be trusted, given the amount of false information he has tweeted in the past.  

On Friday night, prosecutors leveled new accusations against Bankman-Fried in a letter filed to Judge Lewis Kaplan, alleging that Bankman-Fried had been in direct communication with the current general counsel of FTX US, who they called Witness-1 and referred to Ryne Miller, a likely witness at Bankman-Frieds trial in October.  

On Jan. 15, the prosecutors said, Bankman-Fried contacted Miller over the encrypted messaging platform Signal, as well as email, and wrote, I would really love to reconnect and see if theres a way for us to have a constructive relationship, use each other as resources when possible, or at least vet things with each other. 

The prosecutors added that Bankman-Fried had contacted other current and former FTX employees, reflecting what they described as a history of using the application for obstructive purposes, suggesting that Bankman-Fried is trying to influence potential witness testimony.  

Because of Bankman-Frieds activity, prosecutors asked Judge Kaplan for two new bail conditions: that Bankman-Fried not be able to contact current and former employees of FTX or Alameda, and that he not be allowed to use any encrypted or ephemeral call or messaging platforms, such as Signal. Bankman-Fried is currently allowed full access to the internet, which is allegedly one of the reasons he agreed to be extradited back to the United States from the Bahamas.  

The prosecutors argued that the new conditions were necessary to prevent obstruction of justice.  

In a letter filed to Judge Kaplan on Saturday morning, Bankman-Frieds lawyers pushed back, writing that the government was sandbagging any efforts for the two teams to work together on modifying bail conditions.  

According to Bankman-Frieds lawyers, prosecutors had told Bankman-Frieds defense team that they were aware he had contacted John Ray, the new CEO of FTX, as well as Miller on Jan. 19, which the prosecutors viewed as inappropriate. They told Bankman-Frieds lawyers that they would seek an additional bail condition that would prevent Bankman-Fried from contacting potential witnesses without the presence of lawyers.  

Bankman-Frieds lawyers argued that their client had been trying to publicly reach out to Ray to offer help with the bankruptcy and that he had the same charitable motives for reaching out to Miller, describing it as an innocuous attempt to offer assistance.

Even so, they agreed to the proposed bail condition to avoid litigation.

Now, however, his lawyers are arguing that the prosecutors updated restrictionswhich would include contact with all current or former employees, as well as limiting Bankman-Frieds use of messaging platformsare beyond the initial proposed scope.  

The Government apparently, believes that a one-sided presentationspun to put our client in the worst possible lightis the best way to get the outcome it seeks, Bankman-Frieds lawyers wrote, arguing that by filing the letter at 6:00 pm on a Friday, the defense team would not be able to offer a response to the negotiation.  

Bankman-Frieds lawyers proposed a different condition, which would prohibit Bankman-Fried from contacting several key FTX figures, including Caroline Ellison, Zixiao Gary Wang, Nishad Singh, and Witness-1, or Ryne Miller, as well as others to be discussed with the government. They also do not want his access to messaging platforms to be limited.

Finally, they asked Judge Kaplan to remove the bail condition imposed in January that prohibited Bankman-Fried from accessing FTX and Alameda assets, arguing that prosecutors had not offered any evidence that Bankman-Fried was behind the transfers from late December.

Bankman-Frieds lawyers did not immediately reply to a request for comment from Fortune. 

The decision is now in the hands of Judge Kaplan.  

Learn how to navigate and strengthen trust in your business with The Trust Factor, a weekly newsletter examining what leaders need to succeed. Sign up here.

Fast-forward to Monday, and Goldman Sachs has already tossed out that forecast. In a paper bluntly titled 2023 Housing Outlook: Finding a Trough, Goldman Sachs researchers argued the prospects for the U.S. housing market are improving and that national home prices would fall just 2.6% in 2023.

Home sales appear set to turn higher. Mortgage purchase applications have averaged 9% above their October trough so far in January and survey-based measures of purchasing intentions have rebounded sharply, wrote Goldman Sachs researchers in their latest report.

By the time national home prices bottom, Goldman Sachs expects prices to be 6% below their June 2022 peak. Thats also a sharp revision from its previous prediction of a 10% peak-to-trough decline.

This upward revision by Goldman Sachs, coupled with improved mortgage purchase application figures for January, is why Fortune once again tracked down the latest housing forecasts from 29 of the nations leading housing researchers. Among those forecasters, five expect U.S. home prices to grow or remain flat in 2023, while the other 24 believe U.S. home prices have further to fall. (Heres the prior round-up).

Lets take a closer look at the predictions.

Realtor.com: The economics team at Realtor.com predicts that the median price of existing homes will rise 5.4% in 2023 while mortgage rates average 7.4%.

Home.LLC: The firm predicts U.S. home prices will rise 4% in 2023.

CoreLogic: The real-estate data firm predicts that U.S. home prices will rise 2.8% between November 2022 and November 2023. (Here is CoreLogic's latest risk assessment for the nation's 392 largest regional housing markets.)

NAR: The trade group projects that existing home prices are poised to rise 1.2% in 2023 while mortgage rates will average 6.3%.

Freddie Mac: The firm's forecast model has U.S. home prices falling 0.2% in 2023 while mortgage rates average 6.4%. "We expect house prices to decline modestly, but the downside risks are elevated," write Freddie Mac economists.

Mortgage Bankers Association: The trade group's latest forecast has U.S. home prices, as measured by the FHFA US House Price Index, falling 0.6% in 2023 and another 1.2% dip in 2024. The group also forecasts average mortgage rates of 5.2% in 2023 and 4.4% in 2024. "While we would still characterize the path for the national home price index as flat, we are now forecasting several quarters of year-over-year declines in the level of national home prices. We had already been expecting some pretty significant declines in the West and Mountain regions of the country," write researchers at the Mortgage Bankers Association.

Zillow: Economists at the home listing site forecast that U.S. home values will fall 1.1% from November 2022 to November 2023. "In the standoff between buyers and sellers, neither side has fully capitulated, but more sellers have been accepting lower offers," write Zillow economists.

Fannie Mae: Economists at the firm predict that U.S. home prices, as measured by the Fannie Mae HPI, will fall 1.5% in 2023 and another 1.4% dip in 2024. Fannie Mae is currently modeling an average 30-year fixed mortgage rate of 6.3% in 2023 and 5.6% in 2024.

Redfin: The firm's baseline forecast predicts that the median U.S. home sales price will fall 4% in 2023. "Prices would fall more if not for a lack of homes for sale: We expect new listings to continue declining through most of next year, keeping total inventory near historic lows and preventing prices from plummeting," writes Redfin.

HousingWire: Logan Mohtashami, HousingWire's lead analyst, forecast that U.S. house prices in 2023 will see a decline between -5.9% to -7.4% if mortgage rates remain above 5.875%.

Amherst: The real estate investment firm, which owns a massive portfolio of single-family homes, tells Fortune that its forecast model has U.S. home prices falling 5% between September 2022 and September 2025. Incomes are the other side of the seesaw from mortgage rates in setting home prices. We saw no middle-income wage gains for decades. Now its happening big-time. Higher rates are a headwind, but rising incomes are a huge support and tailwind," Sean Dobson, CEO of the Amherst Group, tells Fortune.

Wells Fargo: The bank's forecast model has U.S. home prices falling 5.5% in 2023. "Markets where home prices shot the highest are now vulnerable to a disproportionate swing to the downside, notably in previously white-hot markets in the Mountain West which saw an influx of remote workers at the onset of the pandemic. Home prices in desirable locations with comparatively tighter supply are likely to hold up much better," write Wells Fargo researchers.

Goldman Sachs: On Monday, the investment bank revised its U.S. housing market outlook upwards. It now expects U.S. home prices to fall just 2.6% in 2023. That'd take us, Goldman Sachs says, to a 6% peak-to-trough decline. "We expect a peak-to-trough decline in national home prices of roughly 6% and for prices to stop declining around mid-year. On a regional basis, we project larger declines across the Pacific Coast and Southwest regionswhich have seen the largest increases in inventory on averageand more modest declines across the Mid-Atlantic and Midwestwhich have maintained greater affordability over the past couple years," writes Goldman Sachs researchers.

Capital Economics: Peak to trough, the firm's forecast model has U.S. home prices falling 8%.

Bank of America: The investment bank expects U.S. home prices to decline by around 10% peak-to-trough. "We expect a home price correction (down 10%) rather than a crash (down 15-20%)," wrote researchers at BofA in their Jan. 11 report.

ING: Peak to trough, the Dutch bank tells Fortune it expects U.S. home prices to fall between 5% to 10%. However, the multinational lender says U.S. home prices could possibly decline by as much as 20%. "The housing market downturn, triggered by rapid increases in mortgage borrowing costs, continues to cause us significant concern. Prices have risen hugely over the past couple of years as demand vastly outstripped limited supply of homes, but this process is going into sharp reverse with mortgage applications for home purchases falling by nearly 50% on the 3Q 2020 peak. At the same time there is more supply appearing on the market and the risk we see a steep correction in prices," writes James Knightley, chief international economist at ING.

Bill McBride: McBride, a housing analyst and author of the Calculated Risk blog, expects U.S. house prices to fall by around 10% from the 2022 peak. "Since national house prices increased very quickly during the pandemicup over 40%it seems likely that some of the usual stickiness will not apply. I think the most likely scenario now is nominal house prices declining 10% or more from the peak, and real [adjusted for inflation] house prices declining 25% or so over the next five to seven years," writes McBride.

Keller Williams Realty: Ruben Gonzalez, the chief economist at Keller Williams Realty, expects median home prices as tracked by NAR to fall 10% from top to bottom. "I suspect we will see the trough in the first half of [2023]. Perhaps March or April but that is dependent on the path of interest rates which have been quite volatile recently," Gonzalez tells Fortune.

TD Bank: The bank is forecasting that U.S. home prices will fall by around 10% from peak to trough. That includes a 5.7% drop in 2023 and another 2% drip in 2024. "We see home price growth finding firmer footing soon after the start of 2024," Admir Kolaj, an economist at TD Bank, tells Fortune.

Morgan Stanley: The Wall Street bank expects home prices to fall by around 10% between June 2022 and the bottom in 2024. If mortgage rates fall by more than expected, Morgan Stanley researchers say that the peak-to-trough decline will come in closer to 5%. However, if a "deep" recession manifests, Morgan Stanley predicts U.S. home prices could crash 20% from top to bottom, including up to an 8% home price decline in 2023 alone.

Moody's Analytics: The firm expects a peak-to-trough U.S. home price decline of 10%. If a recession were to manifest, Moody's would expect a top-to-bottom home-price drop of 15% to 20%. This is necessary to re-establish housing affordability, given expected weak household income growth (even without a recession) and mortgage rates that average 6.5% this year before moderating back to 5.5% by mid-decade. If mortgage rates turn out to be lower and/or household income growth stronger, then the house price declines wont be as severe. I have revised the length of time it will take for house prices to reach their trough, which will likely be 2-3 years from now," Mark Zandi, chief economist at Moodys Analytics, tells Fortune.

Zelman & Associates: Back in the summer, the boutique housing research firm forecasted that U.S. home prices would fall 4% in 2023 and another 5% in 2024. According to the Wall Street Journal, the firm now expects U.S. home prices to fall 12% between the 2022 top and the 2024 bottom.

Zonda Home: Peak-to-trough, the real estate analytics firm tells Fortune that its forecast model foresees U.S. home prices falling 15%. "There has been an uptick in traffic (can change that to say buyer interest if better for your audience) since the beginning of the year related to three key things: seasonality, acceptance, and discounts...We still expect prices to be down in 2023 compared to 2022, but how deep of a decline will depend on: how quickly sellers find the market with price cuts, what happens with mortgage rates, how inventory levels trend, and what happens related to a US economic recession," Ali Wolf, chief economist at Zonda, tells Fortune.

AEI: Ed Pinto, director of AEI Housing Center, tells Fortune that he expects U.S. home prices to fall 15% to 20% from peak to trough. Pinto expects prices to bottom out in 2023 or 2024.

CoStar: Peak to trough, CoStar CEO Andy Florance expects U.S. home prices will fall by around 20%. "People who think a 10% drop [in home prices] are dreaming 20% is more comfortable," Florance tells Fortune.

KPMG: The Big Four accounting firm expects U.S. home prices, as measured by the Case-Shiller home price index, to fall 20% between the fourth quarter of 2022 and the fourth quarter of 2023. "The interesting thing to me is how quickly some of these markets are correcting with still very tight inventories Once you start the process of prices falling nationally, there is a self-fulfilling momentum to it, because no one wants to catch a falling knife," Diane Swonk, chief economist at KPMG, tells Fortune.

John Burns Real Estate Consulting: Peak to trough, the real estate research firm's revised forecast has U.S. home prices falling 20% to 22%. That forecast is based on the assumption that mortgage rates stay relatively close to 6% through 2023. "Investors accounted for the highest percentage of buyers ever this cycle in many markets. Lion's share of those [investor] buyers are now on the sidelines, with some needing to sell given overleveraged and really were just taking a flyer on home price appreciation continuing to rip higher. Those days are now over, and these sellers don't exhibit the same emotional and behavioral qualities associated with traditional owner-occupiers, which historically keeps home prices somewhat sticky on the downside," Rick Palacios Jr, the firm's director of research, tells Fortune.

Yieldstreet: By the third quarter of 2023, Yieldstreet expects U.S. home prices to be 20% below its 2022 peak. "Markets that'll be more impacted are the ones where you have a lot of inventory of new homes, like Phoenix, Las Vegas, Dallas, and Boise. There will be some markets in the Northeast, which haven't had a lot of new construction, where home prices are expected to fare better in terms of declines," Tejas Joshi, director of single-family residential at Yieldstreet, tells Fortune.

Pantheon Macroeconomics: The firm expects U.S. existing home prices to fall by around 20%.

Want to stay updated on the housing downturn? Follow me on Twitter at @NewsLambert.